Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Understanding how advanced parametric design can improve the constructability of building designs Shahrokhi, Hooman 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_september_shahrokhi_hooman.pdf [ 33.68MB ]
Metadata
JSON: 24-1.0308669.json
JSON-LD: 24-1.0308669-ld.json
RDF/XML (Pretty): 24-1.0308669-rdf.xml
RDF/JSON: 24-1.0308669-rdf.json
Turtle: 24-1.0308669-turtle.txt
N-Triples: 24-1.0308669-rdf-ntriples.txt
Original Record: 24-1.0308669-source.json
Full Text
24-1.0308669-fulltext.txt
Citation
24-1.0308669.ris

Full Text

UNDERSTANDING HOW ADVANCED PARAMETRICDESIGN CAN IMPROVE THE CONSTRUCTABILITY OFBUILDING DESIGNSbyHooman ShahrokhiB.A., The University of California, Berkeley, 2009A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Civil Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2016c©Hooman Shahrokhi, 2016AbstractConstruction project complexity is rapidly growing, increasing the need for better projectand design management practices. This complexity is attributable to the advancementsin technology which have resulted in an intensification of contemporary design generation.Projects such as the Guggenheim Museum in Bilbao, Spain and the Disney Concert Hallin Los Angeles, California by Frank Gehry & Associates exemplify the results of such tech-nological advancements. This thesis presents the results of a year-long research project inthe form of a case study on a building facade system. The case study approach is chosendue to the exploratory nature of this research project in answering a ‘how’ question wherewe have the ability to observe the events under investigation but lack control over them.The objective of this research is to understand how advanced parametric design can improveconstructability of building designs.The research project was initiated as a test of the hypothesis by our industry partner,an international architecture firm, that they have been able to significantly improve con-structability of a building facade system by using advanced parametric design tools. Whilethe results of this study supported this assertion, a number of recurring challenges werenoticed in the construction process. As a result, phase II of the project was introducedto include a rigorous productivity study of the construction process, identification of delaytypes, categorization of those delay types into constructability issues, and quantifying theimpact of those constructability issues on the project. Finally, a conceptual design approachis developed and used to address the constructability issue with the largest impact by in-corporating construction knowledge - as design rules - into the design stage using advancediiparametric design tools in order to produce a design improved for constructability. Theresults are then validated through a comparison with the original design and checked withan expert subcontractor for confirmation.The claimed contribution of this research project is an increased understanding of howadvanced parametric design tools can be used to improve constructability of the buildingfacade system studied through design.iiiPrefaceFigures 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 3.2, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9 are used withpermission from applicable sources.Chapter 3 serves as a standalone research paper intended for publication as Shahrokhi,H., Staub-French, S., Zadeh, P. A., Poirier, E. A., Diaz, S. I wrote the manuscript with thedirect supervision and guidance of Dr. Staub-French and input from other coauthers.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xivDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv1 Thesis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.2.1 Building Constructability . . . . . . . . . . . . . . . . . . . . . . . . 41.2.2 Parametric Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.2.3 Lean Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.3 Advanced Parametric Design . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.4 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91.5 Research Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10v1.5.1 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.5.2 Phase I Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . 111.5.3 Phase II Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . 131.6 Research Project Context and Scope . . . . . . . . . . . . . . . . . . . . . . 141.7 Thesis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.2 Project Facts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.3 Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.4 Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.5 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 Understanding How Advanced Parametric Design Can Improve the Con-structability of Building Designs: A Case Study . . . . . . . . . . . . . . . 363.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.2.1 Constructability in Design . . . . . . . . . . . . . . . . . . . . . . . . 393.2.2 Advanced Parametric Design . . . . . . . . . . . . . . . . . . . . . . . 403.2.3 Lean Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.3 Research Objectives and Methodology . . . . . . . . . . . . . . . . . . . . . 423.3.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463.3.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493.4 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53vi3.4.1 Phase I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553.4.2 Phase II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 873.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 924.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 924.2 Limitation and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . 94Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95viiList of Tables1.1 Research task contributions toward sub-objectives. . . . . . . . . . . . . . . . 103.1 Comparison of total precast concrete facade panels and those within the scopeof this research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 613.2 Comparison of actual versus planned precast concrete facade panel workthrough 4D simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 653.3 Production cycle delay sampling. . . . . . . . . . . . . . . . . . . . . . . . . 693.4 Method productivity delay model processing. . . . . . . . . . . . . . . . . . . 703.5 Delay information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723.6 Comparison of ideal versus overall productivity. . . . . . . . . . . . . . . . . 723.7 Comparison of panel characteristics. . . . . . . . . . . . . . . . . . . . . . . . 813.8 Sensitivity analysis of the 15-minute time adjustment in some production cycles. 88viiiList of Figures2.1 Progression of architectural design expression. c©August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 192.2 Screenshot of architect’s entire advanced parametric design definition. Adaptedfrom Computational Design in Construction by Diaz, Santiago, February 17,2016. Copyright 2016 by Diaz, Santiago. . . . . . . . . . . . . . . . . . . . . 202.3 Screenshot of architect’s panel type advanced parametric design definitionshowing the 18 different panel types. Adapted from Computational Design inConstruction by Diaz, Santiago, February 17, 2016. Copyright 2016 by Diaz,Santiago. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.4 Complete architectural drawing of precast concrete facade panels includedin IFC package. c© August 2016 by Perkins+Will Canada. Reprinted withpermission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.5 Dimension information for all 18 panel types. c© August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 232.6 Generic panel drawing in different views. c© August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 24ix2.7 Panel type placement on one sample tower elevation. c© August 2016 byPerkins+Will Canada. Reprinted with permission. . . . . . . . . . . . . . . . 252.8 Precast concrete facade panel types PA-01 and PA-01-1 shown side-by-sidedemonstrating the variations. c© August 2016 by APS Precast. Reprintedwith permission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272.9 One of 18 stationary panel formworks on top of a shake table. . . . . . . . . 302.10 Panel formwork with outer wythe steel mesh reinforcement installed. . . . . 302.11 Crew pouring concrete into panel formwork using gantry crane. . . . . . . . 302.12 Crew installing temporary formwork for inner wythe as well as insulation boards. 312.13 Inner wythe steel mesh reinforcement and panel connection hardware installedinto formwork. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.14 Panels arrive on site on an A-frame trailer. . . . . . . . . . . . . . . . . . . . 342.15 Crew installing L-shaped brackets onto the slab. . . . . . . . . . . . . . . . . 342.16 Crew installing connection rods in the back of the panel. . . . . . . . . . . . 342.17 Panel flown into location. Crew receiving the panel at the installation location. 352.18 Crew installing a panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.19 Crew guiding placement of panel and detaching crane hooks from panel. . . . 353.1 Research roadmap showing the relationship of all research tasks and findingsto each sub-objective. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.2 Birds-eye view rendering of case study project showing South Elevation ofNorth Tower and West Elevation of South Tower. c©August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 45x3.3 Process diagram expressing our conceptual approach to detailed design ofprecast concrete facade panels. . . . . . . . . . . . . . . . . . . . . . . . . . . 543.4 Architectural design iterations. c© August 2016 by Perkins+Will Canada.Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563.5 Sample panel type placement on one of four tower elevations. c© August 2016by Perkins+Will Canada. Reprinted with permission. . . . . . . . . . . . . . 573.6 Generic panel drawing from different views. c© August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 583.7 Dimension information for all 18 panel types. c© August 2016 by Perkins+WillCanada. Reprinted with permission. . . . . . . . . . . . . . . . . . . . . . . 593.8 Screenshot of architect’s advanced parametric entire design definition. Adaptedfrom Computational Design in Construction by Diaz, Santiago, February 17,2016. Copyright 2016 by Diaz, Santiago. . . . . . . . . . . . . . . . . . . . . 603.9 Screenshot of architect’s panel type advanced parametric design definition.Adapted from Computational Design in Construction by Diaz, Santiago, Febru-ary 17, 2016. Copyright 2016 by Diaz, Santiago. . . . . . . . . . . . . . . . . 603.10 Count of panels installed per day per week starting on week 1 through week 35. 643.11 Comparison of ideal versus overall production cycle variability. . . . . . . . . 713.12 Two examples of the connection misalignment constructability issue. Exampleb also demonstrates lack of clearance between connection point and structuralcolumn indicated by the yellow circle. . . . . . . . . . . . . . . . . . . . . . . 743.13 Top view of a precast concrete facade panel showing the location and type oflift connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75xi3.14 Close-up of panel connections. . . . . . . . . . . . . . . . . . . . . . . . . . . 783.15 Water collected in a puddle underneath tools. . . . . . . . . . . . . . . . . . 793.16 Panel family type shown with defined variables. . . . . . . . . . . . . . . . . 823.17 Screenshot of proposed advanced parametric design definition that considersthe placement of panel connections according to a set of rules. . . . . . . . . 833.18 Screenshot of a section of the proposed advanced parametric design definitiondemonstrating the rules defined for placement of panel connections based ona range of values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 843.19 Part of a section view through the building looking outward demonstratingthe placement of panel connections with proper clearances from structuralelements. The contents of the red square signify a sample panel and the redcircles point out the panel connections. . . . . . . . . . . . . . . . . . . . . . 853.20 Comparison of a complete set (18) precast concrete facade panels overlaid inplan view of panel manufacturer’s design (top) versus outcome of our concep-tual design approach (bottom). . . . . . . . . . . . . . . . . . . . . . . . . . 86xiiList of AbbreviationsAEC Architecture Engineering ConstructionBIM Building Information ModelCAD Computer-aided DesignIFC Issued for ConstructionMPDM Method Productivity Delay ModelUBC University of British ColumbiaVPL Visual Programming LanguagexiiiAcknowledgmentsThis research project would not have been possible without the help of a number of amazingpeople. First and foremost, I would like to thank my research supervisor, Dr. Sheryl Staub-French, for her support and guidance since day 1 of this journey. I will always be gratefulfor all that I have learned from her. I would also like to thank the members of the BIMTOPiCS Lab and other fellow students - our Postdoctoral Fellows for their guidance, and mycolleagues for their academic and more importantly, moral support. I also wish to expressa very special thank you to Santiago Diaz of Perkins+Will Canada. Without his knowledgeand dedication to this project, this work would not have been complete. Also noteworthy,are the folks at RiteTech Construction for their engagement on this research project; I thankyou. I would also like to thank Dr. Alan Russell for his thoughtful feedback on my workand his kindness throughout the years.Finally, it goes without saying that not only this thesis, but my life would not bewhole without the support of my family, my friends, and last but certainly not least, thelove of my life. I thank each and everyone of you.xivDedicationI dedicate this thesis to those dear to me who left all too early, those close to me who madeit possible for me to be me, and to the one Special, Amazing, Remarkable, and Astoundingperson whom I’m ecstatic to share the rest of this life with.xvChapter 1Thesis Overview1.1 IntroductionConstruction project complexity is increasing rapidly (Gidado, 1996) with many reasonsattributed to this growth. The idea of project complexity has been studied by many scholars(Kim and Wilemon, 2003) and defined in many ways. However, there seems to be a commonunderstanding that complex projects are made up of a number of smaller tasks that vary innature and are difficult to understand (Gidado, 1996; Larson and Gobeli, 1989; Tatikondaand Rosenthal, 2000) . As complexity grows, inherently, the need for more advanced projectplanning must also grow in order to keep up with the demands of the industry. One of theways that building projects are becoming more complex is due to technological advancementswhich lead to more complex building designs. As a result, we see the growing focus on newproject delivery methods and project planning tools such as, integrated project delivery andlean construction principles and tools to help mitigate this complexity.1This research project focuses on complexity through building designs and the con-structor’s ability to reason with the problem. Therefore, constructability - defined as “theoptimum use of construction knowledge and experience in planning, design, procurement,and field operations to achieve overall project objectives” by the Construction Industry In-stitute - plays a major role in this exploration.Through collaboration with industry partners, the research team was presented witha building project where the architecture firm utilized advanced parametric design in anattempt to enhance the constructability of the facade design. The research team looksinto the considerations given by the architecture firm specific to this design consisting of acombination of precast concrete panels and window walls. A thorough study of the architect’sadvanced parametric design model is carried out to gain an understanding of how the designwas translated into a an algorithmic problem as well as to identify areas which address theconstructability of the panels consistent with the architect’s design intent. From this task,we learn of how the architect improved their design by reducing the number of unique facadepanels from about 90 to only 18 different types. Following the handover of the facade panelarchitectural design to the panel manufacturer, we start an investigation into the facade paneldetailed design process. We then find that, through a completely manual design process, thepanel manufacturer has produced 101 different panel variations. While the panel shapes andsizes fit within the 18 different types, the detailed design of the connections can vary withineach panel type depending on its installation location. The panel fabrication process is thenobserved to gain an understanding of the process and coordination of which panel variationis produced when. A productivity study is then carried out in phase II of the researchproject to identify and analyze any issues in the production. Based on those findings, an2understanding is developed of how advanced technologies can be utilized in order to improveconstructability of building designs. We formalize the constructor’s knowledge as designrules using an advanced parametric design tool to produce a design that aims at furtherimproving constructability.As a measure of the architect’s hypothesis, we compare the actual installation progresswith the planned work using the data collected in the productivity study. While the resultspointed toward confirmation of the hypothesis, a number of recurring delays were witnessedin the installation process leading to the introduction of phase II of this research project.Phase II focuses on a more rigorous productivity study examining the installation processat the task-level in order to identify the witnessed types of delays and to measure theirimpact on the panel installation process. The delays are then categorized into design-specific(constructability issues) and non-design-specific issues. Through working sessions with theexpert panel installation subcontractor, an analysis of the constructability issues is carriedout to identify the cause and potential solution to the issues. Finally, a conceptual designapproach is developed for incorporating the knowledge gained from the constructor as designrules into the advanced parametric design system for producing a design that is furtherimproved to address constructability issues. This design approach is explored for addressingthe constructability issue with the largest impact on the panel installation process and theresults are presented in chapter 3 of this thesis.In the following sections of this chapter, we take a look at a number of ideas directlyrelated to this research project and the previous academic work done in these areas. We thendescribe our understanding of advanced parametric design. Finally, we set out our objectivesand the methodologies employed in this research project.31.2 Literature ReviewThe concepts of constructability, parametric design, rule-based design, and lean constructionare well documented in academic literature. However, there is very limited research thatconnects these areas together. Constructability, at its core, requires the incorporation ofconstruction knowledge in design. In this case study project, a number of the recurringchallenges witnessed can be attributed to the lack of such knowledge in the detailed designof these panels. Therefore, we attempt to formalize the construction knowledge as designrules into design using advanced parametric design tools. In order to do so, we must havea clear understanding of parametric design and the abilities it affords designers. Finally, weexplore the principles of lean construction in guiding the generation of a new design aimedat reducing waste and variability in the production cycle.1.2.1 Building ConstructabilityConstructability, as defined by the Construction Industry Institute (1986), “is the optimumuse of construction knowledge and experience in planning, design, procurement, and fieldoperations to achieve overall project objectives.” This definition is widely adopted acrossthe field (Hanlon and Sanvido, 1995; Fischer and Tatum, 1997; Radtke and Russell, 1993;Raviv et al., 2012; Griffith and Sidwell, 1997; Pulaski and Horman, 2005; Lam et al., 2006;Hartmann and Fischer, 2007; Pulaski and Horman, 2008; Wong et al., 2007; Gransbergand Douglas, 2005; Gibson et al., 1996). Most of the research surrounding constructabilityaddresses the implementation and use of constructability in different stages of constructionprojects (Gransberg and Douglas, 2005; Gibson et al., 1996; Griffith and Sidwell, 1997;4Radtke and Russell, 1993; Wong et al., 2007). However, some scholars point to the linkagebetween constructability knowledge and design decisions (Fischer and Tatum, 1997; Lamet al., 2006) which support the idea of improved constructability by formalizing constructor’sknowledge in parametric design. Lam et al. (2006) goes a step further by putting the onuson designers in taking the lead in enhancing the constructability of their designs by carryingout thorough site investigations, coordinating design artifacts and work, and designing for“standardisation, repetition, safety and ease of construction.” In order to make the casefor constructability implementation, common benefits are expressed as reduced cost, shorterschedule, and better control over the process (Gibson et al., 1996; Griffith and Sidwell, 1997;O’Connor, 1985). As in most other works, this research project takes a narrowly focusedlook at one segment of the project stages (Griffith and Sidwell, 1997).1.2.2 Parametric DesignIn order to link the impacts of parametric design on constructability, we must understandhow parametric design differs from conventional design methods including CAD. Woodbury(2010) describes it as such: “[p]arametric modeling (also known as constraint modeling)introduces a fundamental change: ‘marks’, that is, parts of a design, relate and changetogether in a coordinated way.” He argues that conventional design tools make it easy to starta model by adding parts; however, making coordinated changes to the model is difficult andrequires manual inspection of all related parts. Hence, this is the problem that parametricdesign aims to solve as “the system takes care of keeping the design consistent with the5relationships and thus increases designer ability to explore ideas by reducing the tedium ofrework” (Woodbury, 2010).It is crucial to note that in order to be able to reap such benefits of parametricdesign, the designer must have the willingness and ability to define the proper relationshipsas parametric design yields results according to rules and inputs (Woodbury, 2010; Jabi,2013). Jabi (2013) describes a parametric design script as one made up of “standard parts:a declaration of what the script is and does, variables (think of variables as storage units tostore information), functions (specialized and self-contained algorithms that accept input,act on it and produce output) and interfaces (declarations of what buttons, sliders andcheckboxes to display and how to react to them).” He goes on to describe the benefits gainedfrom the invent of object-oriented programming, where instead of variables and functions,full objects can be declared that contain their own set of variables and functions.Gane and Haymaker (2007) describe their parametric methodology around the follow-ing six concepts: variables (primary drivers of geometric variations), constraints (describingthe allowable range of variables), dependency (related constraints causing a change in a vari-able as a result of a change in another variable), component (a geometric assembly definedby variables and constraints), and PowerCopy (a group of components intended to be usedin a context). For example, in the case study explored in this thesis where there are a largenumber of repeating facade panels, the panel can be modeled as a family with characteristicsthat can be inherited in the parametric design script and if needed, manipulated using rulesdefined by the designer (Jabi, 2013).61.2.3 Lean ConstructionFor a few decades now, there has been a push for extending the idea of lean production fromthe manufacturing industry into the construction industry. Early on in this work, Koskela(1992) points to a list of 11 guiding principles, many of which are directly related to thedesign and work carried out in this research project case study (annotated by an asterisk):1. Reduce the share of non value-adding activities.*2. Increase output value through systematic consideration of customer requirements.3. Reduce variability.*4. Reduce the cycle time.*5. Simplify by minimizing the number of steps, parts and linkages.*6. Increase output flexibility.7. Increase process transparency.8. Focus control on the complete process.9. Build continuous improvement into the process.10. Balance flow improvement with conversion improvement.11. BenchmarkAs you will see in the findings of this paper, the measured waste in the work proves tohave a major impact on the entire production process. The importance of waste reduction7has been the subject of many academic papers offering reasons for why waste reductionhas not been conquered in the construction industry (Bolviken and Koskela, 2016) andconceptual approaches to inform elimination of waste (Formoso et al., 2015). One of thereasons leading to an increase in the volume of non value-adding activities (waste) is identifiedas process variability (Koskela, 1992), signifying a need for its reduction in the productioncycle. Inherently, by reducing waste in the production cycle, we are able to reduce the cycletime. Finally, complexity of a product or process can have a negative impact on production(Koskela, 1992). One of the ways for simplifying processes is to standardize parts of theproduct (Koskela, 1992), which is consistent with what we found in this research project.1.3 Advanced Parametric DesignThis section provides necessary background knowledge regarding advanced parametric de-sign. According to Woodbury (2010), the first computer-aided design (CAD) system, IvanSutherland’s PhD thesis on Sketchpad (1963), was parametric. According to Hoffmann andJoan-Arinyo (2005), every major CAD system has adopted parametric solving as of the1980s. In todays CAD systems, for example Autodesk Revit, we know parametric design asthe window that automatically adjusts its location in the model as its host wall is moved.While parametric design is not a new phenomena, there are a number of new tools thatallow for a much more customizable approach to design solutions by the user through VisualProgramming Language (VPL). VPL is a language that allows users to create programs bycreating connections between different elements that carry out specific actions as opposed tocoding textually. The advantage of VPL is that the lack coding knowledge does not hinder8the user’s ability to create a program using this language. Therefore, we refer to such toolsthat afford the user a deeper reach into functionalities by creating a custom program usingVPL “as advanced parametric design tools” and we refer to the design program developedthrough such tools as an advanced parametric design definition.In this research project, advanced parametric design has been utilized in two parts.First, the architecture firm has used an advanced parametric design tool, namely Grasshop-per, in developing the architectural design of the building facade system. The author of thisdesign definition is an individual with expertise in advanced parametric design and carriesthe title of computational designer. Second, a different advanced parametric design tool,Dynamo, has been used by the author of this thesis in the conceptual design explorationperformed to address some of the constructability issues identified as presented in chapter 3.1.4 Research ObjectivesThe objective of this research project is to develop an understanding of how the use of ad-vanced parametric design tools can improve the constructability of building designs. Througha case study, the team explores the constructability of a building facade system from thedesign all-the-way through fabrication and installation. By working with the architecturefirm, the manufacturing subcontractor, and the installation subcontractor, and taking a deeplook into the artifacts produced by each party, we gain an understanding of how the processunfolded. To achieve the overarching research objective of this research project, we formulatethe following sub-objectives:SO1. Understand architectural design process using parametric design.9SO2. Understand detailed design and panel fabrication process.SO3. Assess constructability of panel installation process.SO4. Examine the possibility of incorporating constructability knowledge into parametricdesign.1.5 Research MethodologyThe research methodology has been designed to address the preceding sub-objectives withinthe following sections. Table 1.1 shows the contribution of the research tasks towards eachsub-objective.Table 1.1: Research task contributions toward sub-objectives.Research Task SO1 SO2 SO3 SO4Case studyPanel design explorationShop drawing reviewInterview with designersManufacturing plant visitPanel installation trackingTask-level observationsMethod Productivity Delay modelCause analysisConceptual design exploration1.5.1 Case StudyIn order to meet the objectives of this research project, the following methodologies areadopted. Firstly, a building project is selected to serve as the case study. The selectedproject is made up of two high-rise student housing towers and an adjacent 4 storey academic10and administrative building. The longitudinal facades on the two high-rise towers providethe necessary information to carry out this research project as outlined in the followingsections. The following four stakeholders are identified as subjects of this research project:1) architectural firm responsible for the project design (hereinafter referred to as “architect”);2) construction management firm in charge of the project construction (hereinafter referredto as “CM”); 3) facade panel manufacturing subcontractor (hereinafter referred to as “panelmanufacturer”); and 4) facade panel installation subcontractor (hereinafter referred to as“panel installer”).1.5.2 Phase I Data CollectionThe design rationale and process was captured through presentations and interviews withthe architect in understanding how the project was realized. A thorough analysis of thecomputational design program was then performed with the program author as part of thedesign outcome analysis. The shop drawings produced by the panel manufacturer wereclosely examined in order to understand the differences between the panel types and varia-tions. After a thorough review of the shop drawings, a semi-structured interview was carriedout with the panel manufacturer’s project team to understand the design process and ra-tionale in developing the shop drawings. The panel production process was observed on aday-long production facility tour. The complete process made up of formwork setup, cyclicalfabrication process, panel storage, and delivery were captured in this research task.11Tracking of the panel installation process began with the first panel being installedthrough general on-site observations of the process. During these observations, the followinginformation was collected:A Tower: the tower on which the panel is being installed onto.B Elevation: the elevation on which the panel is being installed onto.C Number of Panels: total number of panels installations during observation.D Panel Number: a unique panel identifier number developed for the purpose of thisstudy. The unique identifier consists of two numbers separated by a dash. The firstnumber indicates the bay in which the panel was installed onto and the second numberindicates the floor on which the panel was installed onto. A bay in this case is acomplete set of panels running vertically from top to bottom of the building facadeand are numbered from left to right.E Installation Date: the actual date on which the panel(s) was installed.Additionally, the following information was collected on an as-needed basis:A Date: the date of the observationB Location: the location of the observation which in many cases included the uniquepanel identifier.C Observation: a thorough description of the observation.121.5.3 Phase II Data CollectionFollowing the panel installation tracking, it was realized that in order to address sub-objectives 3 and 4, a more rigorous analysis of the production cycle was needed. Therefore,a new data collection plan was devised which included the video recording of the completeproduction cycle (from the time a panel is lifted from the laydown area to the time it iscompletely installed). This research task consisted of a constant video recording of the panellaydown area as well as the panel installation location. The two separate videos were thenstitched side-by-side and the times synced in order to capture the full production cycle. In-stallation process activities were then defined and recorded for each worker in the process.The data collected is then used to create a Method Productivity Delay Model and performa thorough delay analysis. We utilize Adrian and Boyer (1976) Method Productivity DelayModel (MPDM) to gain a better understanding of how various delay types impact the pro-duction cycle. Using this method, we are able to generalize the impact of the identified delaytypes on the entire project by comparing the actual productivity of the panel installationtask with the ideal productivity.The cause analysis is directly influenced by the findings from the previous sectionexaminations. In this research task, the identified delay issues were analyzed and catego-rized as constructability or non-design-specific issues. In working with the panel installers,constructability issues were identified and examined more closely in an attempt to formalizethe issue as one that can be expressed in terms of design rules. With the help of an expertcomputational designer, we authored a new advanced parametric design definition using Dy-namo which addresses one of the identified constructability issues. The knowledge gained13from previous research tasks is used as controls and the architectural design as an input inthis new advanced parametric design definition. The new advanced parametric design sys-tem is able to manipulate the parameters which are defined in order to meet the conditionsof the rules set out.1.6 Research Project Context and ScopeIt is crucial to set out the scope of this research project as well as to point out some of thelimitations faced in the process. The four facades clad in the subject panels are made up ofa total of 1,205 panels. The majority of these panels have an almost identical relationshipwith the building grid and structural elements. This study is focused on those 1,086 panels.On the other hand, 119 (under 10%) of the panels require a special connection design dueto their positioning on the structure. These 119 panels have been excluded from the scopeof this research due to sparsity of the collected production cycle data. However, in theory,the same design approach can be applied to the connection design of the excluded panels asthey share the same inputs and controls.1.7 Thesis OverviewThis thesis consists of four chapters; chapter 1 provides an overview of the research project asa whole by setting out the problem, a review of the literature related to the topics discussed,defining a set of objectives, explaining the research tasks carried out, and setting the scope ofthis research project. Chapter 2 provides the necessary building project context information14as related to the case study project. Chapter 3 contains the main body of this researchthesis and serves as the main contribution. This chapter is a standalone research paperwhich addresses the main objective of this research project. Chapter 3 is coauthored bymembers of the research team and an expert computational designer from the case studyproject. Chapter 4 serves as a conclusion to this research project and sets out future works.15Chapter 2Case Study2.1 IntroductionThe research strategy chosen for this project is that of a case study as we attempt to createan understanding of ‘how’ advanced parametric design can improve the constructability ofbuilding designs. According to previous research, case study is favored when asking a ‘how’question and also when the possibility to observe the events under study and to interviewthe persons involved in the events exists but the investigator has little to no control over theevents of this research project (Yin, 2003). Additionally, case studies are particularly usefulwhen the current body of knowledge is limited, providing an in-depth understanding of theresearch subject (Bonoma, 1985; Harris and Ogbonna, 2002; Fellows and Liu, 2015). As partof this case study, we collect quantitative, as well as, qualitative data to address the researchquestion. While the project remains as the medium for the study, a variety of data collectionmethods are carried out in order to facilitate an in-depth understanding of the researchquestion (Yin, 2003; Fellows and Liu, 2015). In particular, interviews of persons involved in16the project are coupled with quantitative data collected during the production process, toform a thorough and generalized understanding of the project production cycle (Fellows andLiu, 2015). In order to achieve a theoretical generalization, we perform a random samplingof the project production cycles to be representative of the entire project (Flyvbjerg, 2006).Finally, we bring the case study to closure by having developed an understanding of howadvanced parametric design can improve the constructability of building designs throughanalysis of the collected data.2.2 Project FactsThe case study focused on the construction of a new building project at the University ofBritish Columbia (UBC) Vancouver campus. The project consists of two student residenthigh-rise towers and an adjacent low-rise academic and administrative building. In total,the project consists of 448,686 ft2 of gross floor area. The student resident towers will house1,049 beds which account for about 75% of the total floor area and the dining/events, teach-ing, and administrative spaces accounting for 7%, 7%, and 4% respectively. The project isscheduled for completion in the Summer of 2016 with a total budget of CA$127.5M. Theproject aims to achieve LEED Gold certification. For the purposes of this research project,we focus on the longitudinal facades of the two towers which consists of a combination ofinsulated precast concrete panels and window walls. Specifically, we are interested in themanufacturing and installation of the precast concrete facade panels. The parties of interestinvolved in this project are: the Architect, responsible only for architectural design of thepanels; the Construction Manager, responsible for coordination of work; the Panel Man-17ufacturer, responsible for panel shop drawings, manufacturing, delivery, and coordinationbetween the Construction Manager and the Panel installer; and lastly, the Panel Installer,responsible for installation of panels.2.3 DesignAccording to the UBC Campus and Community Planning (2010), character objectives arelaid out “to rediscover and accentuate UBC’s unique sense of place and the natural westcoast beauty on the Vancouver campus, to improve the cohesiveness of buildings and land-scapes...” In designing the facades, the Architect establishes the following four themes asdesign guidelines: cultural diversity, nature, reinforce social space, truth in materials. Thedesign is further inspired by seaweeds, leaves, calligraphy, and the use of concrete in express-ing the nature of the material. From this study, was born a facade design that consisted ofa very large number of randomly shaped and sized panels, the cost of which to build, wouldfall outside of the budget.At this point, the design team employs the help of their computational designer inorder to express the architect’s vision in engineering terms that the industry partners canreason with. According to Project Description submitted by the Architect as part of theDevelopment Permit Package, “the individual panels are optimized to minimize the numberof unique panels while still maintaining variety throughout.” The ‘ribbons’ of panels areborn out of a “single line with opposing forces”, representing the unique cultures of thetower inhabitants. The same single line is repeated but shifted vertically in an attempt to“create diversity without adding more complexity” (Figure 2.1).18Figure 2.1: Progression of architectural design expression. c© August 2016 by Perkins+WillCanada. Reprinted with permission.A screenshot of the architect’s advanced parametric design definition is provided infigure 2.2. Diaz (2016) points out the three main pieces of the definition that form the designsolution titled as: 1) calligraphy; 2) modularity; 3) panel types. Calligraphy performs muchof the work as described in the previous paragraphs of defining the curvature, multiplyingit across the surface, and dictating the vertical shift in the panels. Modularity divides theribbons into individual panels at each floor level. Finally, panel types assigns a type to eachpanel and color-coordinates them as shown in figure 2.3.The computational program allowed for an almost instant manipulation of the de-sign using a number of parameters. For example, the pattern in which the ‘ribbons’ shiftvertically, or the amplitude of curvature in the ‘ribbons’, etc. After weeks of consultationwith the project stakeholders, the Architect finalized the design explorations and the fa-cade design was documented in the Issued For Construction (IFC) package on one ARCHE sheet. This one-sheet drawing was the only design document handed over to the panelmanufacturer (figure 2.4).191. Calligraphy3. Panel Types2. ModularityVariablesFigure 2.2: Screenshot of architect’s entire advanced parametric design definition. Adaptedfrom Computational Design in Construction by Diaz, Santiago, February 17, 2016. Copyright2016 by Diaz, Santiago.20Figure 2.3: Screenshot of architect’s panel type advanced parametric design definition show-ing the 18 different panel types. Adapted from Computational Design in Construction byDiaz, Santiago, February 17, 2016. Copyright 2016 by Diaz, Santiago.21OUTER WYTHEPANEL TYPE dim_A dim_B dim_C dim_D dim_E dim_FPA-01 549 mm 1056 mm 2031 mm 649 mm 1150 mm 2113 mmPA-02 649 mm 1150 mm 2113 mm 731 mm 1238 mm 2213 mmPA-03 731 mm 1238 mm 2213 mm 784 mm 1308 mm 2319 mmPA-04 784 mm 1308 mm 2319 mm 794 mm 1351 mm 2425 mmPA-05 794 mm 1351 mm 2425 mm 756 mm 1358 mm 2525 mmPA-06 756 mm 1358 mm 2525 mm 675 mm 1332 mm 2607 mmPA-07 675 mm 1332 mm 2607 mm 566 mm 1277 mm 2660 mmPA-08 566 mm 1277 mm 2660 mm 443 mm 1198 mm 2670 mmPA-09 443 mm 1198 mm 2670 mm 320 mm 1103 mm 2632 mmPA-10 320 mm 1103 mm 2632 mm 211 mm 1004 mm 2551 mmPA-11 211 mm 1004 mm 2551 mm 130 mm 913 mm 2442 mmPA-12 130 mm 913 mm 2442 mm 92 mm 847 mm 2319 mmPA-13 92 mm 847 mm 2319 mm 102 mm 813 mm 2196 mmPA-14 102 mm 813 mm 2196 mm 155 mm 811 mm 2087 mmPA-15 155 mm 811 mm 2087 mm 237 mm 840 mm 2006 mmPA-16 237 mm 840 mm 2006 mm 337 mm 893 mm 1968 mmPA-17 337 mm 893 mm 1968 mm 443 mm 967 mm 1978 mmPA-18 443 mm 967 mm 1978 mm 549 mm 1056 mm 2031 mmINNER WYTHEPANEL TYPE dim_G dim_H WidthPA-01 725 mm 1925 mm 1200 mmPA-02 825 mm 2025 mm 1200 mmPA-03 875 mm 2125 mm 1250 mmPA-04 875 mm 2225 mm 1350 mmPA-05 875 mm 2325 mm 1450 mmPA-06 825 mm 2425 mm 1600 mmPA-07 775 mm 2525 mm 1750 mmPA-08 675 mm 2575 mm 1900 mmPA-09 525 mm 2525 mm 2000 mmPA-10 425 mm 2475 mm 2050 mmPA-11 275 mm 2375 mm 2100 mmPA-12 225 mm 2225 mm 2000 mmPA-13 175 mm 2125 mm 1950 mmPA-14 225 mm 1975 mm 1750 mmPA-15 325 mm 1925 mm 1600 mmPA-16 425 mm 1875 mm 1450 mmPA-17 525 mm 1875 mm 1350 mmPA-18 625 mm 1875 mm 1250 mmAABBFRONT VIEW (ELEVATION) OF PANELSCALE 1:10TOP VIEW OF PANELBOTTOM VIEW OF PANELDATUM LINEGRID LINE (CENTERLINE OF INT. WALL BEYOND)987654321181716151413121110987654321181716151413121110987654321181716151413121110131211101718115161445678923141513181617101112321987645328954181716167111014131512131412101115161723645118789765161598341218171314111012111012131718114151623465897765489321161817151413121110101112131416171815123456987895431181517162121114131076121310111415166589234171817141311101217161578645911823345618116171315141211102987106978453265215184 12131117163215165713121617710984 1567114 151831714181618 151411121318117 14101513164985610781645 171732141211781110918761131410112899 31214131312151698111210151417181 16141510111213181723173162564 1871432548769871015111665131411121415161718171812316151718211413 109761112548763 1710981413151611124 189121 1013811141091312153217183164254 112 911153142165 13 632781514161112141310326789 5461077 3618358 14175 1111215 69 18 15101113141621718168945 11781516172910 4 112137895678111012131568991112131516141711121318143211814161776895415147689101011121315166 347183614 1091213578 2310910131718 1412161512 161717251 1011 87 443 316185698141517 24 112131614157623 1817 1015145113241876324515141617910128431876161579817519 310111413812131421312111213141845 1615117161512610118173109154 18612111851716171151816131496712111215851011121413417161547121121516121110983 1861415141378 513101411 26171812151612 9 310116532981210 321314 1817 1415 98674 1 131216151817 14 5671111843102 171 1615141213111016564123713161211109789131415121317168 21739459818 1211111017 1415181864785126432 173 1817161215166 3 1718 1412133104515149 28187101117181312117616514151615 121314781110214126111013 9965831 1671517211516181817 141314123487910114513 7111516891549 654876 3211012111398589 11832 161713 108651564 1110 71847123 1715791413161514410 414316 13 913 10 149 5 1 16 13 10 4 1 16 13 9 3 14 10 4PRECAST CONCRETE PANELS - TABLE OF VALUES(A COMMON REFERENCE LINEFOR DIMENSIONAL VALUES)FINISHED FLOORUNDERSIDE OF SLABNORTH TOWER SOUTH ELEVATIONNORTH TOWER NORTH ELEVATIONSOUTH TOWER WEST ELEVATION SOUTH TOWER EAST ELEVATIONVIEW A-AAS PERFLASHING &DRAINAGEREQUIREMENTSVIEW B-BAXO VIEW32118161715141213101197481314121011151617236451187891618171211109876543211817161514131211101711121318114151610987564231 dim_D dim_E dim_F dim_A dim_B dim_C50 TYP.125 TYP. dim_G1000   DATUM OFFSET FROM GRID LINE, TYP.25 TYP. INNER WYTHE 'Width'38 TYP.19 x 19 DEEP REVEAL, TYP.75 INSULATION75 INNER WYTHE25 TYP.25 TYP.2410 INNER WYTHE HEIGHT, TYP.2625 OUTER WYTHE HEIGHT, TYP.INNERWYTHEINSULATIONAPPROX. 5°APPROX. 5°TO MATCH TOPCONDITION(SEE VIEW A-A)LEVEL 03LEVEL 04LEVEL 05LEVEL 06LEVEL 07LEVEL 08LEVEL 09LEVEL 10LEVEL 11LEVEL 12LEVEL 13LEVEL 14LEVEL 15LEVEL 16LEVEL 17LEVEL 18LEVEL 19ROOFLEVEL 03LEVEL 04LEVEL 05LEVEL 06LEVEL 07LEVEL 08LEVEL 09LEVEL 10LEVEL 11LEVEL 12LEVEL 13LEVEL 14LEVEL 15LEVEL 16LEVEL 17LEVEL 18LEVEL 19LEVEL 20ROOFNOTE: ROOF LEVEL PRECAST PANELS  TOHAVE OUTER WITHE ONLY SEE 2/A11-92NOTE: PRECAST PANELSTO TERMINATE SEE1/A11-62 AND 4/A11-66Drawing Issue Date080401020SheetTitleApprovedCheckedDrawnJob NumberDateSheet InformationCopyright © 2014 Perkins+WillRevisions1220 Homer St.Vancouver, British ColumbiaCanada V6B 2Y5t 604.684.5446f 604.684.5447www.perkinswill.com16 2345AEBCD16 2345AEBCTENDER ISSUE 2014-07-18ADDENDUM ISSUE 2014-08-01BUILDING PERMIT 2014-10-03ISSUED FOR CONSTRUCTION 2014-10-2010/20/2014 10:51:47 PMc:\temp\411332-UBC PT Orchard Commons- Master_bremnerk.rvtA00-60CLADDING PANELMATRIX(ELEVATION WITHGRID SETOUT)CheckerAuthor411332.0002014.10.20ApproverUBCVANTAGECOLLEGEORCHARDCOMMONSUNIVERSITY OF BRITISHCOLUMBIA, 6363AGRONOMY ROAD,VANCOUVER BC, CANADAUBCPROPERTIESTRUSTNO ISSUE DATEFigure 2.4: Complete architectural drawing of precast concrete facade panels included in IFC package. c© August 2016 byPerkins+Will Canada. Reprinted with permission.22Figure 2.4 is broken into smaller sections (Figures 2.5, 2.6 and 2.7) for legibilitypurposes. The table in figure 2.5 provides the dimensions for the variables shown in thedifferent views for the 18 different panel types. Figure 2.6 shows a generic panel withdefined variables in isometric, top plan, front elevation, and bottom plan views. Figure 2.7demonstrates the placement of different panel types through one sample elevation.OUTER WYTHEPANEL TYPE dim_A dim_B dim_C dim_D dim_E dim_FPA-01 549 mm 1056 mm 2031 mm 649 mm 1150 mm 2113 mmPA-02 649 mm 1150 mm 2113 mm 731 mm 1238 mm 2213 mmPA-03 731 mm 1238 mm 2213 mm 784 mm 1308 mm 2319 mmPA-04 784 mm 1308 mm 2319 mm 794 mm 1351 mm 2425 mmPA-05 794 mm 1351 mm 2425 mm 756 mm 1358 mm 2525 mmPA-06 756 mm 1358 mm 2525 mm 675 mm 1332 mm 2607 mmPA-07 675 mm 1332 mm 2607 mm 566 mm 1277 mm 2660 mmPA-08 566 mm 1277 mm 2660 mm 443 mm 1198 mm 2670 mmPA-09 443 mm 1198 mm 2670 mm 320 mm 1103 mm 2632 mmPA-10 320 mm 1103 mm 2632 mm 211 mm 1004 mm 2551 mmPA-11 211 mm 1004 mm 2551 mm 130 mm 913 mm 2442 mmPA-12 130 mm 913 mm 2442 mm 92 mm 847 mm 2319 mmPA-13 92 mm 847 mm 2319 mm 102 mm 813 mm 2196 mmPA-14 102 mm 813 mm 2196 mm 155 mm 811 mm 2087 mmPA-15 155 mm 811 mm 2087 mm 237 mm 840 mm 2006 mmPA-16 237 mm 840 mm 2006 mm 337 mm 893 mm 1968 mmPA-17 337 mm 893 mm 1968 mm 443 mm 967 mm 1978 mmPA-18 443 mm 967 mm 1978 mm 549 mm 1056 mm 2031 mmINNER WYTHEPANEL TYPE dim_G dim_H WidthPA-01 725 mm 1925 mm 1200 mmPA-02 825 mm 2025 mm 1200 mmPA-03 875 mm 2125 mm 1250 mmPA-04 875 mm 2225 mm 1350 mmPA-05 875 mm 2325 mm 1450 mmPA-06 825 mm 2425 mm 1600 mmPA-07 775 mm 2525 mm 1750 mmPA-08 675 mm 2575 mm 1900 mmPA-09 525 mm 2525 mm 2000 mmPA-10 425 mm 2475 mm 2050 mmPA-11 275 mm 2375 mm 2100 mmPA-12 225 mm 2225 mm 2000 mmPA-13 175 mm 2125 mm 1950 mmPA-14 225 mm 1975 mm 1750 mmPA-15 325 mm 1925 mm 1600 mmPA-16 425 mm 1875 mm 1450 mmPA-17 525 mm 1875 mm 1350 mmPA-18 625 mm 1875 mm 1250 mmPRECAST CONCRETE PANELS - TABLE OF VALUESFigure 2.5: Dimension information for all 18 panel types. c© August 2016 by Perkins+WillCanada. Reprinted with permission.23AABBFRONT VIEW (ELEVATION) OF PANELSCALE 1:10TOP VIEW OF PANELBOTTOM VIEW OF PANELDATUM LINEGRID LINE (CENTERLINE OF INT. WALL BEYOND)(A COMMON REFERENCE LINEFOR DIMENSIONAL VALUES)FINISHED FLOORUNDERSIDE OF SLABVIEW A-AAS PERFLASHING &DRAINAGEREQUIREMENTSVIEW B-BAXO VIEW dim_D dim_E dim_F dim_A dim_B dim_C50 TYP.125 TYP. dim_G1000   DATUM OFFSET FROM GRID LINE, TYP.25 TYP. INNER WYTHE 'Width'38 TYP.19 x 19 DEEP REVEAL, TYP.75 INSULATION75 INNER WYTHE25 TYP.25 TYP.2410 INNER WYTHE HEIGHT, TYP.2625 OUTER WYTHE HEIGHT, TYP.INNERWYTHEINSULATIONAPPROX. 5°APPROX. 5°TO MATCH TOPCONDITION(SEE VIEW A-A)Figure 2.6: Generic panel drawing in different views. c© August 2016 by Perkins+WillCanada. Reprinted with permission.24987654321181716151413121110987654321181716151413121110987654321181716151413121110131211101718115161445678923141513181617101112321987645328954181716167111014131512131412101115161723645118789765161598341218171314111012111012131718114151623465897765489321161817151413121110101112131416171815123456987895431181517162121114131076121310111415166589234171817141311101217161578645911823345618116171315141211102987NORTH TOWER NORTH ELEVATION131412101115161723645118789987654321181716151413121110LEVEL 03LEVEL 04LEVEL 05LEVEL 06LEVEL 07LEVEL 08LEVEL 09LEVEL 10LEVEL 11LEVEL 12LEVEL 13LEVEL 14LEVEL 15LEVEL 16LEVEL 17LEVEL 18LEVEL 19ROOFFigure 2.7: Panel type placement on one sample tower elevation. c© August 2016 byPerkins+Will Canada. Reprinted with permission.252.4 FabricationOf all the explorations performed and logic developed by the design team, the only articlepassed on to the Panel Manufacturer is the drawing shown above in figure 2.4. From thisone sheet, the Manufacturer developed the shop drawings for the 18 different panel types.A sample shop drawing of one of the panel types is shown in figure 2.8. As part of theirwork, the Manufacturer is responsible for the design of panel connections to the structure. Acrucial aspect of the panel connection design is the presence of structural columns and shearwalls, elevator cores, stair wells, and roof configurations adjacent to the back of the panel.For this reason, a panel type PA-03 with equidistant connections from each side of the panelmay work for one location on the facade of the building where there are no conflicts withstructural elements but, may not work in another location where there is a shear wall behindit. Therefore, the manufacturer’s shop drawings show a total of 101 variations to 18 paneltypes. In other words, for the most part, there are 18 different panel shapes and sizes but,there are 101 different panels that might share the same shape and size but have differentpanel connection designs. An example of such panel variation is shown in figure 2.8.26Figure 2.8: Precast concrete facade panel types PA-01 and PA-01-1 shown side-by-sidedemonstrating the variations. c© August 2016 by APS Precast. Reprinted with permis-sion.27From this stage, the Manufacturer proceeds with the fabrication of the panels. A visitto the manufacturing plant reveals the fabrication process as follows:I There are 18 reusable and stationary panel forms in the shape of the panel outer wytheswhich are built on top of shaking tables (Figure 2.9).II Forms are cleaned and oiled prior to the pouring of concrete (Figure 2.9).III Structural mesh is placed inside the form (Figure 2.10).IV Using a gantry crane, a worker positions the concrete bucket over the form and poursenough concrete for the outer wythe while other workers smooth out the concrete(Figure 2.11).V The form is shaken in order to remove air bubbles and final smoothing touches to theedges are made.VI Another set of workers follow along installing the removable formwork for inner wythe(Figure 2.12).VII The insulation is then placed inside the formwork and stakes are inserted into theinsulation that connect the outer and inner wythes (Figure 2.12).VIII Other necessary panel connections, hardware, and structural elements are installedinside the formwork (Figure 2.13).IX The gantry crane is used again to pour the concrete into the formwork while workerssmooth out the concrete and it is shaken.28X After the concrete has set, the panels will be removed from the forms and stored inthe yard.Compared to the installation process, the fabrication has a much faster productioncycle and due to the Manufacturer’s ability to store large number of panels in their yard,panel production did not follow the installation order; rather, as many panels were producedas possible and stored, often for long periods of time, until needed at the project site. Theproject coordinator was in charge of ensuring the correct panel type variation is fabricatedbased on the shop drawings.29Figure 2.9: One of 18 stationary panel formworks on top of a shake table.Figure 2.10: Panel formwork with outer wythe steel mesh reinforcement installed.Figure 2.11: Crew pouring concrete into panel formwork using gantry crane.30Figure 2.12: Crew installing temporary formwork for inner wythe as well as insulation boards.Figure 2.13: Inner wythe steel mesh reinforcement and panel connection hardware installedinto formwork.312.5 InstallationThrough coordination with the Construction Manager and the Panel Installer, the PanelManufacturer would send shipments of 8 to 12 panels at a time on a trailer with an A-frame.The panels usually arrived before or on the day of the installation. At times, more than onetrailer carrying the panels would be stored on site. For logistical reasons, the Panel Installerswere not allowed to use a mobile crane for the panel installations and had to rely on sharingthe tower cranes with the whole project. The tower cranes were owned and operated by theformwork subcontractor and their use was coordinated through the Construction Manager.The panel installation process was as follows:I Panels arrive on site (Figure 2.14).II Installation crew installs L-shaped connection hardware on slab ahead of installationday. Each panel requires four points of connection at the time of installation; twopermanent connections on the bottom and two temporary connections on the top(Figure 2.15).III Panel is attached to crane (Figure 2.16).IV Panel is lifted off the A-frame trailer but kept in laydown area so crew can installconnection rods to the back of the panel at the laydown area (Figure 2.16).V Panel is flown to installation location (Figure 2.17).VI Installation crew receives the panel by inserting the rods into L-shaped connectionhardware and bolting it into position (Figure 2.17).32VII Measurements are made to ensure proper installation (Figure 2.18).VIII Crane hooks are detached and crane goes for the next panel (Figure 2.19).IX Installation crew moves onto next panel installation location.While this concludes the panel installation production cycle as this research projectis concerned, the panel installation process is not fully complete until days later, whena different crew welds the top of the panel to the bottom side of the top floor slab intwo locations and following that the two temporary top connections are removed. Thepreceding scope of work is selected as the subject of this research project. Due to reasonssuch as, accessibility to data or data collection ability, certain parts of the panel designand installation process have been excluded from this scope of work which will be furtherexplained in the next chapter.33Figure 2.14: Panels arrive on site on an A-frame trailer.Figure 2.15: Crew installing L-shaped brackets onto the slab.Figure 2.16: Crew installing connection rods in the back of the panel.34Figure 2.17: Panel flown into location. Crew receiving the panel at the installation location.Figure 2.18: Crew installing a panel.Figure 2.19: Crew guiding placement of panel and detaching crane hooks from panel.35Chapter 3Understanding How AdvancedParametric Design Can Improve theConstructability of Building Designs:A Case Study3.1 IntroductionBuilding designs have become more complex as a result of technological advances in theArchitecture, Engineering, and Construction (AEC) domain, challenging the industry toinnovate in designing complex buildings that are buildable. Cooke (2013) attributes thecomplexity of construction projects to the advancements in technology which have resultedin an intensification of contemporary design generation that has exacerbated the potential36complexity of construction projects. Projects such as the Guggenheim Museum in Bilbao,Spain and the Disney Concert Hall in Los Angeles, California by Frank Gehry & Associatesexemplify the results of such technological advancements in the industry. There is a generalconsensus that the continuous demand for speed, quality, cost control, and technologicaladvancements lead to an increase in complexity which then increases the need for complexproject management objectives (Gidado, 1996; Cooke, 2013; Dubois and Gadde, 2002). Inthe case of the Guggenheim Museum in Bilbao, Spain, the structure was built without the useof any tape measures; structural components were marked with barcodes during fabricationand scanned on site using programs linked to the 3D model which would provide necessaryinformation for their proper installation (LeCuyer, 1997).With the advent of powerful parametric design tools, AEC professionals are able toreconcile building designs made up of contradictory and complex goals and constraints suchas aesthetics, construction costs, program requirements, and building performance (Ganeand Haymaker, 2007). Using advanced parametric design tools, designers can create a sys-tem that generates a design which meets design constraints and allows designers to exploremany options through variations of parameters (Gane and Haymaker, 2007). According toWoodbury (2010), parametric modeling or constraint modeling fundamentally differs fromnon-parametric modeling in that “parts of a design, relate and change together in a coordi-nated way.” In simple terms, parametric design yields a result according to rules and inputs(Jabi, 2013). The architecture domain is seeing a rapid growth in the use of parametricdesign for realization of complex facilities.On the other hand, the topic of poor productivity in the construction industry isthe subject of many academic discussions. As design complexity increases in construction37projects, the issue of constructability becomes more important in addressing the industry’spoor productivity performance (IPENZ, 2008). Constructability as defined by the Construc-tion Industry Institute “is the optimum use of construction knowledge and experience inplanning, design, procurement, and field operations to achieve overall project objectives”(Construction Industry Institute, 1986). By improving constructability in design, the in-dustry can benefit from higher productivity as a result of decreases in delays, decreases inmanpower needed to complete tasks, and decreases in activity duration (O’Connor, 1985).At the core of constructability lies the optimum use of construction knowledge. There-fore, incorporation of construction knowledge becomes integral to the process of improvingconstructability in building design.In this research project, we set out to investigate how the use of advanced parametricdesign tools can improve constructability of building designs. The research project began asa test of a hypothesis put out by an industry partner which claims to have significantly im-proved the constructability of a facade system made up of precast concrete and window wallsusing advanced parametric design tools. By performing a thorough analysis of the architec-tural and detailed design of the precast concrete panels, examination of panel fabrication,and measuring the productivity of panel installations, we find that the precast concrete panelproduction is carried out as planned which is significantly improved compared to the ini-tial design proposal which did not benefit from the use of advanced parametric design tools.However, during the panel installation process, we observed a number of recurring challengeswhich seemed to cause delays in the production cycle. As a result, we added an additionallayer to the research project aimed at exploring the challenges witnessed by performing arigorous task-level productivity study which lead to the identification of a number of delay38types. We categorize these delay types into constructability issues and non-design-specificdelays. We then explore a conceptual approach to solving one of the constructability issuesby incorporating construction knowledge gained through consultation with the subcontractorin charge of panel installations into a new parametric design definition developed in consulta-tion with the architect’s computational design expert. The results of this study, as presentedin this paper, suggest an improvement in building design constructability when constructionknowledge is used in developing a rule-based design system with advanced parametric designtools as the means for manipulating the design.3.2 Background3.2.1 Constructability in DesignConstructability, as defined by the Construction Industry Institute (1986), “is the optimumuse of construction knowledge and experience in planning, design, procurement, and fieldoperations to achieve overall project objectives.” While most of the research surroundingconstructability addresses the implementation and use of constructability in different stagesof construction projects (Gransberg and Douglas, 2005; Gibson et al., 1996; Griffith andSidwell, 1997; Radtke and Russell, 1993; Wong et al., 2007), some scholars point to the linkagebetween constructability knowledge and design decisions (Fischer and Tatum, 1997; Lamet al., 2006) which support the idea of improved constructability by formalizing constructor’sknowledge in advanced parametric design. In order to make the case for constructabilityimplementation, common benefits of its use are expressed as reduced cost, shorter schedule,39and better control over the process (Gibson et al., 1996; Griffith and Sidwell, 1997; O’Connor,1985).3.2.2 Advanced Parametric DesignOne of the most impactful advancements in AEC technology tools has been the adventof parametric design. Woodbury (2010) describes it as such: “[p]arametric modeling (alsoknown as constraint modeling) introduces a fundamental change: ‘marks’, that is, parts of adesign, relate and change together in a coordinated way.” He argues that conventional designtools make it easy to start a model by adding parts; however, making coordinated changesto the model is difficult and requires manual inspection of all related parts. Hence, this isthe problem that parametric design aims to solve as “the system takes care of keeping thedesign consistent with the relationships and thus increases designer ability to explore ideasby reducing the tedium of rework” (Woodbury, 2010). It is crucial to note that in order tobe able to reap such benefits of parametric design, the designer must have the willingnessand ability to define the proper relationships as parametric design yields results accordingto rules and inputs (Woodbury, 2010; Jabi, 2013). Jabi (2013) describes a parametric designscript as one made up of “standard parts: a declaration of what the script is and does,variables (think of variables as storage units to store information), functions (specialized andself-contained algorithms that accept input, act on it and produce output) and interfaces(declarations of what buttons, sliders and check boxes to display and how to react to them).”Gane and Haymaker (2007) describe their parametric methodology around the following sixconcepts: variables (primary drivers of geometric variations), constraints (describing the40allowable range of variables), dependency (related constraints causing a change in a variableas a result of a change in another variable), component (a geometric assembly defined byvariables and constraints), and PowerCopy (a group of components intended to be used ina context).3.2.3 Lean ConstructionFor a few decades now, there has been a push for extending the idea of lean production fromthe manufacturing industry into the construction industry. Early on in this work, Koskela(1992) points to a list of 11 guiding principles, many of which are directly related to thedesign and work carried out in this research project case study (annotated by an asterisk):1. Reduce the share of non value-adding activities.*2. Increase output value through systematic consideration of customer requirements.3. Reduce variability.*4. Reduce the cycle time.*5. Simplify by minimizing the number of steps, parts and linkages.*6. Increase output flexibility.7. Increase process transparency.8. Focus control on the complete process.9. Build continuous improvement into the process.4110. Balance flow improvement with conversion improvement.11. BenchmarkAs you will see in the findings of this paper, the measured waste in the work proves tohave a major impact on the entire production process. The importance of waste reductionhas been the subject of many academic papers offering reasons for why waste reductionhas not been conquered in the construction industry (Bolviken and Koskela, 2016) andconceptual approaches to inform elimination of waste (Formoso et al., 2015). One of thereasons leading to an increase in the volume of non value-adding activities (waste) is identifiedas process variability (Koskela, 1992), signifying a need for its reduction in the productioncycle. Inherently, by reducing waste in the production cycle, we are able to reduce the cycletime. Finally, complexity of a product or process can have a negative impact on production(Koskela, 1992). One of the ways for simplifying processes is to standardize parts (Koskela,1992) directly correlating with our findings.3.3 Research Objectives and MethodologyWe set out to test a hypothesis by our industry partner, an international architecture firm,that they have been able to significantly improve the constructability of a facade system- as part of their complete architectural design of a student housing project in Vancouver,British Columbia - by translating their facade design into an algorithmic problem usingadvanced parametric design tools and producing a rule-based design. We, therefore, posethe following research question: how can constructability of building designs be improved byusing advanced parametric design tools? We initiated a case study to investigate how the use42of advanced parametric design tools can improve constructability of building designs. Thecase study approach was selected due to the exploratory nature of this research project inanswering a ‘how’ question where we have the ability to observe the events under investigationbut lack control over them (Yin, 2003).The overarching objective of this research project leads to the development of thefollowing sub-objectives. Figure 3.1 shows a roadmap for this research project divided intotwo phases as described in the following sections. The first level after the sub-objectives infigure 3.1, identifies the research tasks and the following level identifies the findings fromeach research task.SO1. Understand architectural design process using parametric design.SO2. Understand detailed design and panel fabrication process.SO3. Assess constructability of panel installation process.SO4. Examine the possibility of incorporating constructability knowledge into parametricdesign.43Figure 3.1: Research roadmap showing the relationship of all research tasks and findings to each sub-objective.44The subject of this case study is a building project on the University of BritishColumbia (UBC) Vancouver campus. The project consists of two student resident high-risetowers and an adjacent low-rise academic and administrative building. In total, the projectconsists of 448,686 ft2 of gross floor area. The towers stand just under 150 meters tall andspan 66 meters along the widest elevation. The facade system consists of a combinationof insulated precast concrete panels and window walls. The focus of this research projectis the precast concrete panels on the two wider elevations of the high-rise towers visible infigure 3.2 and their opposite facades.Figure 3.2: Birds-eye view rendering of case study project showing South Elevation of NorthTower and West Elevation of South Tower. c© August 2016 by Perkins+Will Canada.Reprinted with permission.We are specifically interested in the manufacturing and installation of the precastconcrete panels. There are a total of 1,205 panels on 4 elevations, of which, 119 panels havebeen excluded from the scope of this research due to sparsity of the collected production cycledata. These 119 panels consist of highly unique panel variations designed for installationat the structural cores and the roof level which cannot be generalized with the rest of the45panels. The parties of interest involved in this project are: the architect, responsible only forarchitectural design of the panels; the construction manager, responsible for coordination ofwork; the panel manufacturer, responsible for panel shop drawings, manufacturing, delivery,and coordination between the construction manager and the panel installer; and lastly, thepanel installer, responsible for installation of panels.3.3.1 Data Collection3.3.1.1 Phase IAs previously noted, the initial goal of this research project was to test the hypothesis thatconstructability of the facade system has been improved using advanced parametric designtools. In order to test this hypothesis, a number of research tasks were carried out. Aspart of sub-objective 1, understand the architectural design process, most importantly theconstructability efforts in design, we analyzed the parametric design definition authored bythe architecture firm and interviewed the project architect, facade architect, and computa-tional designer. The design rationale and process was captured through presentations andinterviews with the project architect. The advanced parametric design definition was ex-plored through multiple working sessions with the computational designer. A step-by-stepanalysis of the different components of the computational design definition was carried outwith the author. The nature of visual programming language - where variables, constraints,dependencies, and operators form components of the design logic - allows for a progressionalpursuit of the logic in the design definition. On the other hand, the extent of parametricmodeling of the precast concrete panels ends with the architect. As part of sub-objective462 - understand the detailed design and panel fabrication process - we reviewed the shopdrawings produced by the panel manufacturer, conducted a semi-structured interview witha number of their project team members, and observed the complete cycle of producing oneset of panels (18 different types).Finally, to meet sub-objective 3 - assess the productivity of the panel installationprocess - we track panel installations from the beginning of the project. The observationswere done in three main ways: 1) on-site observations of panel laydown areas; 2) on-siteobservations of panel installation location; 3) off-site observations through time-lapse imagesfrom two different locations showing: a) North elevation on North tower; b) South elevationof North tower; c) West elevation on South tower. Through these observations, the number ofpanels installed per day and their specific location was gathered for the entire project. Whilethe initial findings pointed toward confirmation of our hypothesis, through comparison of theactual panel installation versus the planned panel installation, we noticed a set of recurringchallenges that were causing delay in the installation production cycle. This was a pivotalmoment in the research project which lead to the introduction of a new phase focused atidentifying and measuring the amount of delays resulting from the observed constructabilitychallenges and the exploration of a potential solution.3.3.1.2 Phase IIWhile the above mentioned observations and tracking of panel installations per day wereuseful in analyzing the panel installation system as a whole, the data lacked the detailsnecessary for a rigorous analysis of the panel installation process, identification of delay types,and relevant constructability issues that impacted productivity. As a result, an additional47data collection method was introduced aimed at capturing the production cycle in its entiretyas part of sub-objective 3 - assess constructability of panel installation process. This newdata collection method consisted of installation of video cameras at two locations: 1) panellaydown area; 2) panel installation location. These cameras captured a wide-angle view ofthe panel lifting process at the laydown area and the delivery of the panel to the work surfaceand the entire panel installation process. By utilizing this data collection method, we areable to perform a productivity study at the task level which enables us to identify delay typesin the production cycle and measure the amount of delays attributable to these identifiedissues. We utilize Adrian and Boyer (1976) Method Productivity Delay Model (MPDM) togain a better understanding of how various delay types impact the production cycle. Usingthis method, we are able to generalize the impact of the identified delay types on the entireproject by comparing the actual productivity of the panel installation task with the idealproductivity.Following the identification of the delay types, an analysis is carried out to establishthe cause and potential solution to each of the issues. Through this research task, the iden-tified delay types are categorized as constructability issues or non-design-specific delays. Asthe final stage of this research project, we explore a conceptual framework for incorporat-ing construction knowledge into design using advanced parametric design tools. In order toaddress the constructability issue using parametric design tools, we must be able to expressthe solution in terms of parameters, constraints, and relationships between the elements ofdesign. We develop a new parametric design model in an attempt to express a solution to oneof the identified constructability issues in terms of characteristics of the panels with whichthe design can associate rules and constraints that embody engineering significance (Shah48and Mntyl, 1995). This proposed design goes beyond the architect’s attempt at improvingconstructability in reducing the number of panel types by incorporating the constructor’sknowledge into addressing the most impactful constructability issue observed. This explo-ration has been performed in direct consultation with the architect’s computational designerand the resulting design is examined by the panel installation subcontractor for validity.Additionally, the design outcome of our conceptual approach is compared with the panelmanufacturer’s design in order to visualize the improvements gained.3.3.2 Data Analysis3.3.2.1 Phase IAs a first step in exploring the architectural design process, we spoke with the facade ar-chitect to gain an understanding of the design rationale and objectives in order to see howthey were translated into the parametric design definition. In working closely with the com-putational designer, we examine the parametric design definition in detail over six workingsessions. The progressional nature of the visual programming language allows for a step-by-step examination of the entire definition. During this research task, we analyze the definitionby looking for areas where rules and constraints are used to manipulate the design and seetheir effect on the design. Finally, the resulting design outcome, in the form of a single-page,which is passed on to the panel manufacturer as part of the IFC drawings is examined forthe information and amount of detail that it entails.Next, a thorough analysis of the shop drawings prepared by the panel manufacturerwas carried out to understand the detailed design of each panel type. During this task, we49noticed of a range of panel variations which also becomes a subject of this investigation.The examination of the panel variations leads to a semi-structured interview with the panelmanufacturer’s project team. The goal of this research task was to understand the rationaleand process behind the detailed design which resulted in a large number of panel variations.In addition, during a plant visit, we examine the way in which panels are manufactured,stored, and delivered to the project site. Particularly, we were interested in the order inwhich panels are manufactured as the design and on-site installations require a randomselection of panels for each installation session as the horizontal order in which they areinstalled does not allow for a standard selection of panels.While the qualitative data described thus far gave us a detailed understanding of thedesign and manufacturing of the panels, the test of whether the panel design improved con-structability could not yet be answered. In order to answer this question, we start trackingthe panel installations on site to compare with the planned activities. The planned activitiesare established as a test of this hypothesis as it was determined to be a significant improve-ment from the initial architectural design - which did not benefit from the use of advancedparametric design - during the conceptual design stage by the architect and constructionmanager and that it was comparable to other traditional facade systems. Thus, we set outto track the installation of panels by collecting the date and specific location of each panelinstalled. This information is then compared to planned panel installation activities agreedupon by the subcontractors and the construction manager. A side-by-side 4D model of theactual versus planned panel installation is created to better visualize and compare the panelinstallation process. During this research task, a number of recurring challenges were noted50in the panel installation process which lead to the introduction of phase II of this researchproject.3.3.2.2 Phase IIIn this phase of the research project, we sample and analyze the complete installation processfrom the time a panel is being attached to the crane at the laydown area to the time it iscompletely installed and the crew is ready to relocate to the next panel. We define thiscomplete process as the production cycle. The production cycle for 31 panels has beencaptured over 28 hours of video recordings from two sources - laydown area and panelinstallation location. Through repeated observations of the production cycle, a complete listof tasks performed by all workers and tower crane is created to facilitate the data collection.For each production cycle, the time per task per worker (including crane) is collected inorder to identify the delay areas and their magnitude to be used in the Method ProductivityDelay Model (MPDM).The method productivity delay model (MPDM) relies on the collection of data per-taining to the production cycle as well as the nature of delays during the cycle (Dozzi et al.,1993) which best fits the needs of this research project in identifying the waste areas. UsingMPDM we can identify the sources of delays and their relative contribution to the productioncycle productivity. This technique allows for measuring the production as well as informingdecisions on how to improve the productivity as the delay information provides a thoroughunderstanding of the impact of the different delay types on the production cycle (Adrianand Boyer, 1976). The first step in the process is to have a sample of production cycleswhich we collected for a total of 31 production cycles. Using the collected data, we were able51to process the method productivity delay model in order to link the collected data and therelative measure of risk and variability in the production cycle (Adrian and Boyer, 1976).Using this information, we are able to calculate and compare the ideal cycle variability andthe overall cycle variability for the observed production cycles. Using the sampled data, weare able to measure the number of each delay occurrence and the total time it adds to theproduction cycle, which in turn, allows us to calculate the probability of each delay typeoccurrence, its relative severity, and the expected percentage of delay time per productioncycle. Furthermore, we compare the productivity of nondelayed versus overall productioncycles in terms of units per hour and project the total time spent nondelayed and overallproduction cycles to demonstrate the total time lost due to the identified delay issues.The final research task in addressing sub-objective 3 - assess constructability of panelinstallation process - is to analyze the cause of each of the identified delay types in orderto separate constructability issues from the rest. The findings from this research task willalso inform our design exploration in the following research task. The delay cause analysisis performed through panel installation observations over more than 30 site visits, numerousdiscussions with the panel installation work crew, and 3 working sessions with the panelinstaller’s managing director, project manager, and project foreman over the course of panelinstallations. The primary focus of this task is to: 1) separate the delays caused by con-structability issues from others; and 2) inform the development of a rule-based solution inour design exploration. It is important to separate non-design-specific delay types as theireffect on productivity of the panel installation process is not related to lack of constructabil-ity in design. Similarly, any potential improvements in design would not affect those delaytypes.52Once there has been a case made for potential improvements in the production cycle,we explore a design approach using an advanced parametric design tool, specifically Dynamo,as a mechanism for developing an advanced parametric design system controlled by a setof rules - identified through the previous research task - which could potentially furtherimprove the design to address the identified constructability issue with the largest measuredimpact on the production cycle. The advanced parametric design definition was developedin direct consultation with the architect’s computational designer as an expert in currentbest practices. This task is aimed at addressing sub-objective 4 - examining the possibilityof incorporating constructability knowledge into parametric design. The design rules weredeveloped through direct consultation with the panel installers as previously described. Theresulting panel design is checked against the panel manufacturer’s design to compare theextent of panel type variations and also the causes for the identified constructability issues.We also validate the improvements in design through the panel installers as an expert withfirst-hand knowledge of the challenges experienced due to the identified constructabilityissues.3.4 FindingsIn answering our research question - how can constructability of building designs be improvedby using advanced parametric design tools? - the exploratory nature of this case studyaffords us the ability to investigate a potential solution to the identified constructabilityissue with the largest impact on the production cycle of panel installations. Through thisinvestigation, we are able to highlight our main finding through the process diagram shown53in figure 3.3. At the center of the diagram lies the function of developing an advancedparametric design system capable of exploring a problem through variations of parameters(Gane and Haymaker, 2007) as described in section Phase II of Findings. This system isdeveloped using an advanced parametric design tool, Dynamo, as the mechanism to achievingthe function. Construction knowledge gained through design examination, observations,productivity studies, and consultations with constructors informs the creation of designrules which act as a control to the function. The necessary project details such as, buildingstructure and panel type shapes, in the form of a BIM serve as inputs to the system. Finally,the system is able to compute a design solution that meets the rules and constraints definedby the designer. In the following sections, we present the findings of all research tasks aslaid out in the methodology section which leads to the generation of the aforementionedconceptual framework.Figure 3.3: Process diagram expressing our conceptual approach to detailed design of precastconcrete facade panels.543.4.1 Phase IThe building facade system, as it was originally designed, was made up of about 90 differentpanel types. One of the design themes, set out early on in the design process, was to expressthe natural characteristics of concrete which can take the shape of any form it is pouredinto. Therefore, it was decided that the facade panels would be made of concrete. Thepanels are made up of three layers in order from interior to exterior: 1) inner wythe: 3-inchthick concrete slab that is the will be finished to form the interior wall; 2) 3-inch thicklayer of insulation; 3) outer wythe: 3 to 5-inch thick concrete slab that forms the finishedexterior of the panel. As one can imagine, the cost of fabrication alone, could render thissolution unfeasible. The initial design solution was presented to the construction managerand precast concrete subcontractor for cost feedback which was not approved by the projectowner. The design had to be modified or scrapped entirely to meet budget requirements.At this stage, the architect enlists the help of their computational designer to translate thisdesign solution into an algorithmic problem which could potentially address the fabricationcost issue. The design begins with a straight vertical line; opposing horizontal forces areapplied to each end to create a curve for expressing the fluidity of concrete; the same curveis then copied but shifted vertically to form the other edge of the panels; and finally, theset is copied across the facade of the building while methodically shifted up and down tocreate diversity. The complete set is then cut horizontally at each floor level to create setsof only 18 different panel types running along the height of the building positioned in away to express the original design intents of uniqueness and randomness. The parametricnature of this design solution allowed for an almost-instant generation of design iterations55with the change of a number of variables such as the intensity of curvature in the panels asdemonstrated in figure 3.4.Figure 3.4: Architectural design iterations. c© August 2016 by Perkins+Will Canada.Reprinted with permission.After a period of design consultation within the architectural firm and with the owner,a design solution was selected and presented to the constructors for cost feedback which metthe budget and was comparable to other traditional facade systems. Once approved, thefinal design outcome was produced on one ARCH E sheet included in the IFC drawings andhanded over to the construction manager and eventually the panel manufacturer for detaileddesign. The one page drawing is split into three figures for better legibility. The drawingincludes: 1) 4 elevations showing the position of each panel type (figure 3.5); 2) a genericpanel shape from multiple views (figure 3.6); 3) a table housing the sizing characteristics ofthe 18 panel types (figure 3.7). The extent of parametric design ends at this point in timewhen all the architectural design is handed over to the subcontractor as a one-sheet drawing.56987654321181716151413121110987654321181716151413121110987654321181716151413121110131211101718115161445678923141513181617101112321987645328954181716167111014131512131412101115161723645118789765161598341218171314111012111012131718114151623465897765489321161817151413121110101112131416171815123456987895431181517162121114131076121310111415166589234171817141311101217161578645911823345618116171315141211102987NORTH TOWER NORTH ELEVATION131412101115161723645118789987654321181716151413121110LEVEL 03LEVEL 04LEVEL 05LEVEL 06LEVEL 07LEVEL 08LEVEL 09LEVEL 10LEVEL 11LEVEL 12LEVEL 13LEVEL 14LEVEL 15LEVEL 16LEVEL 17LEVEL 18LEVEL 19ROOFFigure 3.5: Sample panel type placement on one of four tower elevations. c© August 2016by Perkins+Will Canada. Reprinted with permission.57AABBFRONT VIEW (ELEVATION) OF PANELSCALE 1:10TOP VIEW OF PANELBOTTOM VIEW OF PANELDATUM LINEGRID LINE (CENTERLINE OF INT. WALL BEYOND)(A COMMON REFERENCE LINEFOR DIMENSIONAL VALUES)FINISHED FLOORUNDERSIDE OF SLABVIEW A-AAS PERFLASHING &DRAINAGEREQUIREMENTSVIEW B-BAXO VIEW dim_D dim_E dim_F dim_A dim_B dim_C50 TYP.125 TYP. dim_G1000   DATUM OFFSET FROM GRID LINE, TYP.25 TYP. INNER WYTHE 'Width'38 TYP.19 x 19 DEEP REVEAL, TYP.75 INSULATION75 INNER WYTHE25 TYP.25 TYP.2410 INNER WYTHE HEIGHT, TYP.2625 OUTER WYTHE HEIGHT, TYP.INNERWYTHEINSULATIONAPPROX. 5°APPROX. 5°TO MATCH TOPCONDITION(SEE VIEW A-A)Figure 3.6: Generic panel drawing from different views. c© August 2016 by Perkins+WillCanada. Reprinted with permission.58OUTER WYTHEPANEL TYPE dim_A dim_B dim_C dim_D dim_E dim_FPA-01 549 mm 1056 mm 2031 mm 649 mm 1150 mm 2113 mmPA-02 649 mm 1150 mm 2113 mm 731 mm 1238 mm 2213 mmPA-03 731 mm 1238 mm 2213 mm 784 mm 1308 mm 2319 mmPA-04 784 mm 1308 mm 2319 mm 794 mm 1351 mm 2425 mmPA-05 794 mm 1351 mm 2425 mm 756 mm 1358 mm 2525 mmPA-06 756 mm 1358 mm 2525 mm 675 mm 1332 mm 2607 mmPA-07 675 mm 1332 mm 2607 mm 566 mm 1277 mm 2660 mmPA-08 566 mm 1277 mm 2660 mm 443 mm 1198 mm 2670 mmPA-09 443 mm 1198 mm 2670 mm 320 mm 1103 mm 2632 mmPA-10 320 mm 1103 mm 2632 mm 211 mm 1004 mm 2551 mmPA-11 211 mm 1004 mm 2551 mm 130 mm 913 mm 2442 mmPA-12 130 mm 913 mm 2442 mm 92 mm 847 mm 2319 mmPA-13 92 mm 847 mm 2319 mm 102 mm 813 mm 2196 mmPA-14 102 mm 813 mm 2196 mm 155 mm 811 mm 2087 mmPA-15 155 mm 811 mm 2087 mm 237 mm 840 mm 2006 mmPA-16 237 mm 840 mm 2006 mm 337 mm 893 mm 1968 mmPA-17 337 mm 893 mm 1968 mm 443 mm 967 mm 1978 mmPA-18 443 mm 967 mm 1978 mm 549 mm 1056 mm 2031 mmINNER WYTHEPANEL TYPE dim_G dim_H WidthPA-01 725 mm 1925 mm 1200 mmPA-02 825 mm 2025 mm 1200 mmPA-03 875 mm 2125 mm 1250 mmPA-04 875 mm 2225 mm 1350 mmPA-05 875 mm 2325 mm 1450 mmPA-06 825 mm 2425 mm 1600 mmPA-07 775 mm 2525 mm 1750 mmPA-08 675 mm 2575 mm 1900 mmPA-09 525 mm 2525 mm 2000 mmPA-10 425 mm 2475 mm 2050 mmPA-11 275 mm 2375 mm 2100 mmPA-12 225 mm 2225 mm 2000 mmPA-13 175 mm 2125 mm 1950 mmPA-14 225 mm 1975 mm 1750 mmPA-15 325 mm 1925 mm 1600 mmPA-16 425 mm 1875 mm 1450 mmPA-17 525 mm 1875 mm 1350 mmPA-18 625 mm 1875 mm 1250 mmPRECAST CONCRETE PANELS - TABLE OF VALUESFigure 3.7: Dimension information for all 18 panel types. c© August 2016 by Perkins+WillCanada. Reprinted with permission.A screenshot of the architect’s advanced parametric design definition is provided infigure 3.8. Diaz (2016) points out the three main pieces of the definition that form the designsolution titled as: 1) calligraphy; 2) modularity; 3) panel types. Calligraphy performs muchof the work as described in the previous paragraphs of defining the curvature, multiplyingin across the surface, and dictating the vertical shift in the panels. Modularity divides theribbons into individual panels at each floor level. Finally, panel types assigns a type to eachpanel and color-coordinates them as shown in figure 3.9.591. Calligraphy3. Panel Types2. ModularityVariablesFigure 3.8: Screenshot of architect’s advanced parametric entire design definition. Adaptedfrom Computational Design in Construction by Diaz, Santiago, February 17, 2016. Copyright2016 by Diaz, Santiago.Figure 3.9: Screenshot of architect’s panel type advanced parametric design definition.Adapted from Computational Design in Construction by Diaz, Santiago, February 17, 2016.Copyright 2016 by Diaz, Santiago.60From the one sheet drawing, the panel manufacturer produces their detailed designconsisting of 101 panel variations as documented in their shop drawings. While there are only18 different panel types (shape and size), the detailed design results in a subset of differentpanel connection designs which add to the total number of varying panels. For the mostpart, these variations are in the positioning of connection points with the building structure.As previously mentioned, the scope of this research project excludes 119 specific panels -less than 10% of the total 1205 panels - which have a highly unique connection design andthe sparsity of the collected production cycle data pertaining to these specific panels. Thisinformation is summarized in table 3.1.Table 3.1: Comparison of total precast concrete facade panels and those within the scope ofthis research.Panels Type Count Variations Connection Location VariationsWithin Research Scope 18 1,086 38 34Overall 18 1,205 101 34Through a semi-structured interview with the panel manufacturer’s design team fol-lowing a thorough review of the shop drawings, we learned that the designers manuallyproduced a standard panel connection design for the 18 panel types at first. They thengo on to identify areas of conflict between their standard panel connection design and thebuilding structure and manually modify those panel connections. As a result, a number ofthe panel types have multiple variations and therefore, such a large number of panel varia-tions exists. It is important to note that while there may be a large number of variations foreach panel type, the variation does not affect the panel shape and size. Therefore, duringfabrication, a complete set of 18 panels are produced which may be of different variationsaddressed by different configurations of the embedded hardware. During this interview, we61also learn from the chief engineer that the biggest challenge in panel installations is theinaccuracy of embed placements in the concrete slabs which was an early indication of theobserved delay issues discussed further on. During the manufacturing plant visit guided bythe project coordinator, we observe the entire process of fabricating 18 panels from start tofinish. There are 18 fixed and reusable formworks built atop shaking tables for each of the 18different panel types. The first step of the process is to strip the formworks from the panelsthat were constructed the day before and removing the panels from the tables to the storagearea using a gantry crane. The forms are then cleaned and a new set of 18 panels are manu-factured in the following order: 1) placement of outer wythe steel reinforcement; 2) pouringof outer wythe concrete; 3) placement of insulation and connections between insulation andthe two wythes; 4) attachment of the inner wythe formwork according to the specificationsfor each specific panel variation; 5) pouring of inner wythe concrete; 6) the concrete curesovernight. According to the project coordinator, step 4 in the process proves to be the mostchallenging as there is an additional level of complexity created by the large number of panelvariations. Furthermore, as the panels are installed onto the structure floor-by-floor, they donot follow any systematic order. However, a complete set of panels (1 through 18) is manu-factured during each manufacturing cycle to increase efficiency in the workflow. Therefore,a large number of panels are stored at the panel manufacturer’s facility - about 400 at thetime of our site visit - to be shipped to the job site as they are needed. According to theproject coordinator, this task also adds another level of complexity in tracking and knowingwhich panel variation to produce when and where to have it stored for accessibility when itis needed. All of these findings confirm and emphasize our interest in the large number ofpanel variations.62As described in the methodology sections, we had set out to the test the hypothesisthat using advanced parametric design, the architect has been able to improve constructabil-ity of the facade system and as a measure of that test we began tracking panel installationsince the first day of panel installations. The panel installations began on the North tower asit was further ahead in the schedule. The data collected was frequently checked against theplanned work. Early on in the process, it was confirmed that the panel installation processwas in fact not only able to keep up with the scheduled work but it was even faster thanplanned. While the initial agreement between the construction manager and panel installerswas to share the project crane for two half-days during the week and the entire day on theweekends, they modified their agreement to install panels on weekends only until the toweris completed and the crane is not needed by other subcontractors. Panel installations perday per week are plotted on the graph shown in figure 3.10. The blue bars indicate panelinstallations on the North tower while the orange bars represent the South tower. As noted,panel installations during the first two weeks took place during the weekends as well as week-day until a new agreement was reached to carry out installations on the weekends only as aresult of the faster than planned panel installations. Also documented in this graph is theshift in weekend work from the North tower to the South on week 15 and the continuationof North tower work during the weekdays.63Figure 3.10: Count of panels installed per day per week starting on week 1 through week 35.Additionally, we visualized the panel installation process for the North elevation ofthe North tower by creating a 4D simulation of the planned versus actual work to confirmour finding with an additional level data not represented in the previous graph, the panellocation. Table 3.2 summarizes this data for the first four weeks of work. You will noticea number of missing panels on week four which is also confirmed by a lesser number ofinstallations on week four in figure 3.10 due to factors beyond the scope of this researchproject such as issues with the tower crane. However, the simulation clearly demonstratesthe ability to install the panels as planned confirming the hypothesis that the architect isable to improve constructability in design with the use of advanced parametric design tools.Once the initial hypothesis is confirmed, we move on to a detailed investigation the observeddelays in the production cycle and those findings are described in the following subsection.64Table 3.2: Comparison of actual versus planned precast concrete facade panel work through4D simulation.Week Actual Planned1234653.4.2 Phase II3.4.2.1 Task-Level ObservationsDuring the initial on-site observations, the panel installation process was captured and brokendown into a complete set of tasks performed by the work crew and the crane. This sub-categorization of installation tasks allows for a uniform method of capturing the times spentperforming each activity by each worker. The panel installation data collection included:a) workers 2 and 3 at the laydown area responsible for attaching the correct panel to thecrane hook for transportation to the installation location; b) the crane identified as worker1; c) workers 4 and 5 at the installation location responsible for attaching the panel tothe structure. During this research task, we collected detailed information about the timespent per worker per task in the panel installation process. This information confirmed oursuspicion of recurring challenges that delayed the panel installation production cycle. Theseevents are identified as delay types to be used in the method productivity delay model. Ageneral description of each delay type is provided below.I Connection misalignment: this issue occurs when a panel is picked up by the craneand flown into the installation location but the rods in the back of the panel do notalign with the L-shaped connection hardware in the slab. There are multiple reasonswhy this type of issue occurs and will be discussed in the next chapters. This issue isresolved by temporarily installing the panel with one missing connection and returningat a later time to address the fourth connection.II Lift delay: this issue occurs when a panel is lifted off of the A-frame trailer at an angle.The remedy to this issue is to lower the panel back onto the A-frame trailer and one66of the workers will adjust the chain links by hand to ensure a level lift off. At timesthis issue occurs more than once with the same panel as the chain link adjustmentsare only made based on intuition and not measured.III Extensive adjustments: this issue occurs when a panel is placed into location for instal-lation but is slightly off mark. The issue is that the panel is either not vertically levelor that it needs to move horizontally in one direction or another. The vertical levelissue is due to an unlevel floor slab and faulty correction of the slope by placing plasticshims underneath the panel. The horizontal misalignment is caused when the L-shapedconnection hardware is slightly off mark and pulls the panel to one side or another. Inthese cases, the installation crew spends more time than average on the adjustment ofthe panel location but unlike in Issue I, Misalignment, the panel is eventually installedwith all four connections.IV Environmental delay: at times, rainwater would collect in puddles around the worksurface. In such cases, during the ‘bolt’ activity as defined above, the electrical breakerwould go off when power extensions fell into these water puddles. This would cause aloss of power to the electrical bolt drivers and essentially halt the process. A workerwould then have to remove the extension cord from the puddle, dry it off, and resetthe breaker.V Crew transition delay: as previously mentioned, panels were installed floor-by-floor.Once the crew would reach the end of one floor, they would either move across thetower to the other elevation or a floor up to continue the installations. This transitioncould delay their readiness to receive and install the next panel.673.4.2.2 Method Productivity Delay ModelThe necessary data captured captured through complete video recordings of the productioncycle is then populated into the delay sampling table (Table 3.3) as part of the methodproductivity delay model analysis (Adrian and Boyer, 1976). For ease of referencing, theproduction cycles are numbered in the order they were observed but are sorted in this ta-ble by the adjusted production cycle time. The delay column identifies the type of delay:misalignment (M), lift (L), adjustments (A), environmental (E), crew transition (T), nonde-layed (N). As described in the delay type section above, the misalignment issue is resolvedby temporarily installing the panel with one missing connection and addressing the fourthconnection at a later time. In these cases, the installation of the fourth connection happenedoutside of the production cycle captured in the video recordings and near impossible for theresearch team to collect that data due to fast-pace nature of the work and the variability inprocess of going back to address the issue. For this reason, a 15-minute time adjustment isapplied to all production cycles where a misalignment issue occurred. The 15-minute adjust-ment is deemed to be an accurate representation of the time it takes to address the properinstallation of the fourth connection. In production cycles where the delay is caused by morethan one delay type, a percentage is calculated from the recorded times attributed to eachdelay type. Finally, the last column shows the absolute value of the production cycle timeminus the mean value for the nondelayed production cycles and is later used to measure thevariability of production cycle times.68Table 3.3: Production cycle delay sampling.ProductionCycleDelayCycleTimeCycleTime (s)Cycle TimeAdj. (s)ConnectionMisalignmentLiftDelayExtensiveAdjustmentsEnvironmentalDelayCrewTransitionDelayMinus MeanNon-DelayTime (s)Panel 27 L 0:11:11 671 671 0% 100% 0% 0% 0% 225Panel 31 N 0:11:57 717 717 0% 0% 0% 0% 0% 179Panel 2 N 0:12:49 769 769 0% 0% 0% 0% 0% 127Panel 24 N 0:13:08 788 788 0% 0% 0% 0% 0% 108Panel 23 N 0:13:10 790 790 0% 0% 0% 0% 0% 106Panel 28 T 0:14:54 894 894 0% 0% 0% 0% 100% 2Panel 1 N 0:15:00 900 900 0% 0% 0% 0% 0% 4Panel 12 N 0:15:24 924 924 0% 0% 0% 0% 0% 28Panel 20 A 0:16:53 1,013 1,013 0% 0% 100% 0% 0% 117Panel 29 A,L 0:17:13 1,033 1,033 0% 16% 84% 0% 0% 137Panel 18 L 0:17:15 1,035 1,035 0% 100% 0% 0% 0% 139Panel 8 N 0:17:16 1,036 1,036 0% 0% 0% 0% 0% 140Panel 10 N 0:17:38 1,058 1,058 0% 0% 0% 0% 0% 162Panel 15 N 0:18:06 1,086 1,086 0% 0% 0% 0% 0% 190Panel 5 A 0:20:19 1,219 1,219 0% 0% 100% 0% 0% 323Panel 25 A 0:20:28 1,228 1,228 0% 0% 100% 0% 0% 332Panel 26 A,L 0:21:44 1,304 1,304 0% 42% 58% 0% 0% 408Panel 30 A 0:23:14 1,394 1,394 0% 0% 100% 0% 0% 498Panel 17 A 0:24:17 1,457 1,457 0% 0% 100% 0% 0% 561Panel 7 T,M 0:12:33 753 1,653 96% 0% 0% 0% 4% 757Panel 11 M 0:14:28 868 1,768 100% 0% 0% 0% 0% 872Panel 16 T,L 0:32:03 1,923 1,923 0% 18% 0% 0% 82% 1,027Panel 13 M 0:17:07 1,027 1,927 100% 0% 0% 0% 0% 1,031Panel 21 M 0:17:43 1,063 1,963 100% 0% 0% 0% 0% 1,067Panel 6 E 0:33:59 2,039 2,039 0% 0% 0% 100% 0% 1,143Panel 22 M 0:20:14 1,214 2,114 100% 0% 0% 0% 0% 1,218Panel 14 M,A 0:21:45 1,305 2,205 76% 0% 24% 0% 0% 1,309Panel 9 M 0:22:51 1,371 2,271 100% 0% 0% 0% 0% 1,375Panel 19 L 0:37:58 2,278 2,278 0% 100% 0% 0% 0% 1,382Panel 3 L,M 0:25:11 1,511 2,411 66% 34% 0% 0% 0% 1,515Panel 4 L,E,M 0:29:23 1,763 2,663 76% 4% 0% 20% 0% 1,767Total 10:07:11 36,431 44,531 9,612 2,303 1,939 1,496 873 17,491Mean nondelay (s) 89669By performing the MPDM processing, we are able to link the collected data and therelative measure of risk and variability in the production cycle (Adrian and Boyer, 1976).Table 3.4 compares such data for the nondelayed production cycles and the overall productioncycles.Table 3.4: Method productivity delay model processing.UnitsProductionTotal Time (s)Numberof CyclesMean CycleTime (s)∑[|(CycleT ime)−(NondelayCycleT ime)|]nNondelayed production cycles 8,068 9 896 116Overall production cycles 44,531 31 1,436 588Using this information, we are able to calculate and compare the ideal cycle variabilityand the overall cycle variability for the observed panel installations as follows. While thevariability in both cases seems high, in the nondealyed production cycles, it amounts tounder 2 minutes which can be justified as there are a number of varying factors in theproduction cycles, such as, the distance the panel has to travel from the laydown area tothe installation location, the weight and thus maneuverability of the panel, and other siteconditions. However, compared to the ideal cycle variability, the significant increase in theoverall cycle variability should be of concern. The significant disparity between the two isdemonstrated in figure 3.11.Ideal cycle variability∑[|(CycleT ime)−(NondelayCycleT ime)|]÷nMeanCycleT ime= 116896= 0.13Overall cycle variability∑[|(CycleT ime)−(NondelayCycleT ime)|]÷nMeanCycleT ime= 5881436= 0.4170 6001100160021002600Ideal OverallTime (s)Production CycleProduction Cycle VariabilityFigure 3.11: Comparison of ideal versus overall production cycle variability.To further investigate the impact that each delay type has on the production cycle, wecalculate - for each delay type - the information shown in table 3.5. Adrian and Boyer (1976)describe the rows in table 3.5 as follows: a) the ‘probability of occurrences’ is a measure ofdelay cycles per total number of cycles; b) the ‘relative severity’ is a measure of the meanadded cycle time per mean overall cycle time; c) the ‘expected percentage of delay time perproduction cycle’ is calculated by multiplying the previous two values. The relative impactof each delay type on the production cycle is presented in table 3.5. While the number ofoccurrences for the first three delay types is almost equal, it is crucial to note the varyinglevels of impact on the production cycle.71Table 3.5: Delay information.Time VarianceMisalignmentDelayLiftDelayAdjustmentsEnvironmentalDelayTransitionDelayOccurrences 9 8 8 2 3Total added time 9,612 2,303 1,939 1,496 873Probability of occurrences 0.290 0.258 0.258 0.065 0.097Relative severity 0.74 0.20 0.17 0.52 0.20Expected percentage ofdelay time per production cycle21.6% 5.2% 4.4% 3.4% 2.0%We then calculate the productivity of both nondelayed and overall production cycleson an hourly basis. Furthermore, we calculate the total time spent on panel installationsby multiplying the measured data by the total number of panels that were installed in thisbuilding project (Table 3.6). The difference in the total hours (150 hours) is time lost as adirect result of the identified delays. In the following section, we explore, in more detail, thecauses of the identified types.Table 3.6: Comparison of ideal versus overall productivity.Mean CycleTime (s)Productivity(units/hr)TotalHoursNondelayed production cycles 896 4.0 270Overall production cycles 1,436 2.5 4333.4.2.3 Cause AnalysisOnce we know the impact of the identified delay types on the production cycle, we focuson the causes for each of the delay types. As this research project aims to provide an un-derstanding of how advanced parametric design can improve the constructability of buildingdesigns, you will notice a focus on the constructability issues which are deemed to be solvableusing advanced parametric design. We explore these constructability issues and the potentialfor addressing them through the lens of rule-based design.72Connection misalignmentAs previously noted, the design specification for the precast concrete facade panelsrequire four points of connected with the building structure. During the panel installationprocess, there are two permanent connections made on the bottom of the panel and twotemporary connections made at the top. All four connections are made with L-shapedbrackets which are previously installed onto the structure. Before the panel is transportedinto its installation location using the crane, workers install four threaded rods in the back ofthe panel that are designed to match the L-shape brackets on the structure. As demonstratedin the previous section, in a considerable number of production cycles, the rods do not alignwith the connections. When the misalignment is significant enough where the tolerancesprovided in the bracket do not allow for the proper installation of the panel, the panel istemporarily installed with at least 3 connections. At times, the misalignment is identifiedprior to transporting the panel into its installation location, in which case, the respectiverod is not installed on the back of the panel prior to its transport into location. At othertimes, the misalignment only becomes apparent once the panel is transported into location,at which point, the workers are forced to stop the ideal installation process and addressthe misalignment by either removing the rod from the back of the panel or removing theL-shaped bracket installed on the building structure (figure 3.12).Through extensive observations and discussions with the installation crew, the causesfor such event have been identified as: 1) connection design which doesn’t allow enoughclearance between the L-shaped bracket and structural elements such as, columns or shearwalls; 2) connection design which creates a physical clash between the rod in the back of thepanel and structural elements such as, columns or shear walls; and 3) large number of panel73type variations lead to a large number of varying locations for where the L-shaped bracketsmust be installed which increases the chances of errors in connection installation location.Figure 3.12: Two examples of the connection misalignment constructability issue. Example balso demonstrates lack of clearance between connection point and structural column indicatedby the yellow circle.Through successive meetings and discussions with the panel installers as well as thearchitect’s computational designer, it was determined that using rule-based design we couldeliminate the possibility of any physical clashes or inadequate clearances while also improvingthe bracket installation accuracy by reducing the number of varying bracket locations andthus reducing the number of panel type variations. In this case, the following informationare known and available from the BIM: 1) location of structural columns and shear walls; 2)location of panels relative to the grid lines; and 3) size of the panel inner wythes. In theory,using an advanced parametric design tool, this information can be used as inputs and a set74of rules can be defined in order to calculate the location of panel connections which addressesthe issues mentioned in the previous paragraph.Lift delayThe process for transporting each panel from the laydown area to its installationlocation consists of attaching the panel to the crane at two points: 1) flexible steel cableembedded into the inner wythe of the panel (point a in figure 3.13); 2) lifting pin embeddedinto the panel outer wythe (point b in figure 3.13). The panel is then lifted off of the A-frametrailer and held in position while the workers check it for a level lift off and installation ofthe threaded rods in the back of the panel. The panels are required to be transported tothe installation location in a horizontal level because they need to fit in between two floorsand the rotation in the case of an unlevel liftoff causes the raised corner at the top and thelowered corner at the bottom to clash with the building floor slabs.Figure 3.13: Top view of a precast concrete facade panel showing the location and type oflift connections.As demonstrated in the previous section, in a considerable number of occasions, thepanel does is not lifted at a horizontal level and is then placed back onto the truck foradjustments. The lift adjustment is a purely arbitrary activity where the worker in chargeof attaching the crane hooks to the panel will determine to shorten or lengthen a few links75on the chain based on intuition. This process consists of completely lowering the crane mainhook block - about 50-80 ft - and moving the correct chain a few links up or down on themain hook block in order to shorten or lengthen the chain and once again raising the mainhook block and lifting the panel off of the A-frame trailer to be tested again for its level. Attimes, this process would be repeated three to four times. The activity durations marked as‘re-hook’ account for the time spent on this type of adjustment. As shown in figure 3.13, oneof the crane connection points on the panel is a flexible steel cable; the length of which wasarbitrarily adjusted by the fabrication workers as it was discovered during the fabricationsite visit. Furthermore, even if the length of the steel cable was prescribed and maintainedequally throughout all panel fabrications, the positioning of the two hooks does not accountfor the center of gravity of the panels. Therefore, the panel lift becomes a fortuitous taskpurely dependent on the worker’s intuition and the hope that length of the cable steel remainsequal throughout the fabrication.In theory, as the geometry and density of the materials that make up the panelsare known, we should be to calculate the center of gravity of the panel and place the liftconnections accordingly. In this case, the panel geometry and density information would beused as an input in the computational program, a rule would be defined for placing two liftconnections on the top of the panel which exert an equal and opposite force to the forceat the center of gravity. The computational program would then be able to calculate thecorrect location for the panel lift connection.Extensive adjustmentsSimilar to the first two issues, this constructability issue has a high probability ofoccurrence and a significant contribution to delays in the production cycle. As previously76mentioned, there are two contributory factors to this constructability issue. First, wheninstalling the panels, the level of the floor slab is corrected by using plastic shims which thepanel sits on. At times, this adjustment is poorly crafted; therefore, once the panel is placedinto position, it will be out of level. The solution to this issue is to lift the panel only fromthe side that needs adjustment, removing the pack of shims - which consists of a selection ofdifferent size shims - and adjusting the thickness of the shim to account for the level. Theshim pack is then placed back into place and the panel is lowered onto the shim pack. Attimes, this issue would require repeating the corrective tasks in order to obtain the desiredlevel. In such cases, the fault in the process lies strictly with the floor slab being out of levelwhich is well within the standards for concrete floor slabs. For this reason, the solution liesoutside the scope of this work.Second, there are times when the panel is placed into location but requires a horizontalshift in one direction or another. At times, this is caused just by faulty placement of thepanel onto the slab, in which case, the solution is fairly simple. Using special pry bars,the workers shift the panel from one side to another. At other times, this issue is causedwhen the L-shaped bracket is installed slightly off-mark which causes a horizontal shift inthe panel’s location once bolted into place. The solution to this constructability issue isan on-the-go adjustment of the L-shaped bracket. As shown in figure 3.14, the L-shapedbrackets provide a fair amount of tolerance - addressed by the square washer installed priorto the nut - which allows for the adjustment of the rod to obtain the desired installationlocation. Similar to the misalignment issue, this is a case of the threaded rods in the back ofthe panel not properly aligning with their desired installation location. The only differencebeing the extent at which the alignment is off mark. By that logic, the proposed solution77to the misalignment issue should also have a positive and direct impact on the adjustmentsissue.Figure 3.14: Close-up of panel connections.Environmental delayAt times, after or during a rainy day, water would collect in puddles around the panelinstallation location (figure 3.15). This would cause the electrical tools to short the circuitwhen they came into contact with the water. As a result, power to the tools would be lostwhich caused a delay in the ‘bolt’ activity. Depending on the panel installation location, thepower may have been provided using an extension cord from across the building floor. Thiswould cause a significant delay as a worker would then need to walk across the floor to resetthe electrical breaker. As this delay type is not considered a design issue, it will be excludedfrom the scope of this work.78Figure 3.15: Water collected in a puddle underneath tools.Crew transition delayNormally, panel installations start at one end of the floor on one elevation and continuethrough that elevation. Once the row is complete, the crew moves across the width of thefloor to the other elevation and continues to installations in that row. Once the floor iscompleted, the crew transitions to the floor above. At times, this transition caused a delay inthe crew’s readiness to receive and install the next panel. The panel lifting, rod installation,and transportation provides enough time for the crew to move from one installation to thenext adjacent location but, when the crew moves from one floor to another, multiple tripsare made to relocate all tools and materials to the new floor. As this delay type is notconsidered a design issue, it will be excluded from the scope of this work.3.4.2.4 Proposed Advanced Parametric DesignFollowing a thorough analysis and understanding of the panel design, fabrication, and in-stallation process, we explore a new approach to the detailed design process hinted to atthe beginning of the findings section. By developing a new advanced parametric design79definition, we are able to generate a panel connection design which takes into account thelearnings of the previous sections in addressing the connection misalignment and extensivealignment constructability issues. The system focuses on the three identified causes for thisconstructability issue (figure 3.12): 1) lack of clearance between L-shaped bracket and struc-tural elements; 2) physical clash between threaded rod on the back of the panel and structuralelements; and 3) error-prone task of installing L-shaped brackets due to the large numberof varying locations. The four elements around the function in figure 3.3 are described asfollows:I Building information model: the BIM provides us with the geometry of the buildingdesign and the relative placement of panels in relation to the structural elements.II Construction knowledge: through discussions with the panel installers and thoroughon-site observations, it was determined that panel connections required adequate clear-ances from structural elements. Having known the required clearances and geometricalinformation of the panel types, we are able to define a set of rules for manipulating theplacement of panel connections on the back of panels. Using the information availablefrom the BIM, we calculate the range of clearance available from the edge of the panelto the structural obstacle and place the panel connection according to the range ofclearance available. In this method, we are automatically minimizing the placement ofconnections in varying locations as the ranges are calculated relative to the structuralelements.III Advanced parametric design: in this case, Revit add-in, Dynamo, is utilized as amechanism to perform this function. The reason Dynamo is selected is because the80BIM was available to the research team in Revit format and Dynamo is freely availableas an add-in to Revit.IV Constructability in design: finally, the resulting design is one with adequate clearancesfor all connection brackets, avoidance of panel connection clashes with structural ele-ments, and reduction of varying connection designs as indicated in table 3.7.Table 3.7: Comparison of panel characteristics.Panels Type VariationsConnection LocationVariationsOur Proposed Design 18 18 7Original Design Within Research Scope 18 38 34Original Design Outside Research Scope 18 101 34Similar to any other CAD system, the design solution can be set up in a number ofdifferent ways. The following paragraphs describe how we have set up the problem usingAutodesk Revit and Dynamo. First, we create a generic panel as a family type and assign thesame variables as in the architect’s one-sheet drawing to the panel family type. In addition,we have introduced two new variables for the position of the left and right connection pointsoff the gridline labeled as dim LC and dim RC (figure 3.16). While the values for dim Athrough dim H will be imported from the table provided by the architect in the IFC drawings,the values for dim LC and dim RC are parametrically calculated based on the requiredclearances and determined according to the range of possibilities.81dim_Hdim_Gdim_Adim_Bdim_Cdim_Fdim_Edim_Ddim_LCdim_RCGrid LineFigure 3.16: Panel family type shown with defined variables.In a new Autodesk Revit project file, we link the building structure model devel-oped by the engineering firm. The structural information is needed as the calculations willreference this information for placement of the panel family type and adjust connection po-sitioning accordingly. Figure 3.17 shows a screenshot of our Dynamo definition. Grouped inorange, these modules obtain the necessary information from the building structural modeland pass it on to the Panel Family modules for placement of panel family types. The greengroups of modules take the preceding information along with the panel dimension information82obtained from the architect’s one-sheet IFC drawing and imported through a spreadsheet(pink group) to manipulate the dimension variables of the generic panel family type into thedesired values.LocationPanelFamilyDimension Informationdim_LC/RCComputationImportSpreadsheetFigure 3.17: Screenshot of proposed advanced parametric design definition that considersthe placement of panel connections according to a set of rules.In addition to dimensions obtained from the spreadsheet for dim A through dim H,the two blue groups serve as design rules developed based on the previous findings of thisresearch project to to determine the desired placement of panel connections. These rules(figure 3.18) obtain the necessary clearance range information from the preceding modulesand depending on a set of defined ranges, determine the placement of each panel typeconnection.83Figure 3.18: Screenshot of a section of the proposed advanced parametric design definitiondemonstrating the rules defined for placement of panel connections based on a range ofvalues.The resulting design, as shown in table 3.7, reduces the number of panel variationsdown to 18 from 38 while cutting down varying connection locations from 34 to 7. Thereduced number of panel variations has the potential of improving productivity during panelmanufacturing as well as the panel installation process. As we learned from the panel man-ufacturer’s project coordinator, the two most complex aspects of the panel manufacturingmanagement process were the coordination of fabricating the correct panel variation neededand also keeping track of the variations for when they need to be delivered to the job site.Eliminating variations per panel type would eliminate the additional level of complexity inthe panel manufacturing process of coordinating the fabrication of multiple variations as wellas keeping track of those variations for delivery. Additionally, elimination of variations perpanel type eliminates the need for keeping track of variations at the job site and adjusting84work procedures during panel installations as some variations require different setups forinstallation. As for connection clearances and clashes with the structure - the top two iden-tified causes of the connection misalignment constructability issue - because the advancedparametric design system is bound to follow a set of rules that require certain clearances fromthe building structure, the risk of designing connections that conflict with the structure iseliminated. Figure 3.19 shows a snapshot of a section through the building looking outwardsdemonstrating placement of panel connections without conflicts with the structure.Figure 3.19: Part of a section view through the building looking outward demonstratingthe placement of panel connections with proper clearances from structural elements. Thecontents of the red square signify a sample panel and the red circles point out the panelconnections.85The third identified cause of the connection misalignment constructability issue wasdeemed to be the error-prone task of placing L-shaped connection brackets onto the buildingstructure at 34 random locations per set of panels. The advanced parametric design systemincludes a design rule which measures an allowable range for the position of each panel con-nection and determines the exact placement based on a set of defined options to minimize thenumber of varying connection locations. The results of this rule are displayed in figure 3.20in comparison with the panel manufacturer’s design for a standard set of panels. A completeset of panels (1 through 18) are overlaid in plan view to demonstrate the varying locationof panel connections in figure 3.20. In doing so, we have minimized the chances of errorsin placement of the L-shaped brackets onto the floor slabs. The results of this design havebeen checked with the panel installers as experts in the installation process who confirm ourassumption that laying out connections at 7 locations as opposed to 34 could significantlydecrease the chances of errors considering the large number (4,344) of connections that areinstalled in this project.Figure 3.20: Comparison of a complete set (18) precast concrete facade panels overlaid inplan view of panel manufacturer’s design (top) versus outcome of our conceptual designapproach (bottom).863.5 DiscussionThe evaluation of the hypothesis that, the use of advanced parametric design has improvedconstructability of the facade panels, takes into consideration factors outside of the designerscontrol. While the results suggest an improvement in constructability as compared to theinitial design, there are challenges in the larger project sense that affect the panel installersability to carry out their work as planned. These challenges are outside the scope of work ofthe panel installers, manufacturer, and architect as it relates to the facade system. However,the results of these challenges are visible in figure 3.10, specifically in weeks 5, 13, 18-20, orthe last few weeks of installation on the North tower. A few examples of such challengesare: unavailability of crane during its raising process, windy days, serious crane malfunctionfor a long period of time, and unprepared work surface due to delays in predecessor activity.The productivity assessment takes into account the ability to install the number of units perweek as initially planned should all other factors affecting the production also take place asplanned.Also noteworthy, is the adjustment in production cycle time by adding 15 minutesto each cycle where a connection misalignment occurs. While the 15-minute addition is notmeasured, it is deemed to be an accurate representation of the time it takes to address theissue, at a later time, outside of the normal production cycle. This number is determinedbased on direct observations of the production cycle of over 75 panels by our on-site researcherand in consultation with the panel installers. By performing a sensitivity analysis of ±20%on the value 15, we observe the changes presented in table 3.8 which still account for thelarge majority of the impact on the production cycle.87Table 3.8: Sensitivity analysis of the 15-minute time adjustment in some production cycles.Time Adjustment-20% 15 Minutes +20%Relative severity 0.64 0.74 0.84Expected percentage of delay time per production cycle 18.5% 21.6% 24.4%Production (units/hour) 2.6 2.5 2.4In regards to the placement of connection points in our conceptual design, one mightraise the issue of structural engineering requirements. While calculations specific to ourconnection designs have not been carried out, we have taken this issue into considerationin defining the rules for connection placement. The ranges defined for the placement ofpanel connections all fall within the range of varying connection location for each paneltype. For example, if the bottom left connection in the manufacturer’s design for panel type1 is placed between 75 and 250 mm from the edge of the inner wythe in three differentvariations, our design takes this into consideration by defining an allowable range for theconnection placement. Through this measure, we have mitigated the need for validatingstructural design requirements. Should we have had the input of the structural engineer,we may have been able to expand these ranges and allow for further standardization of theconnection placement resulting in less varying connection locations on the floor slabs.As exemplified in this research project, the architecture firm is the only party withthe resources and capabilities to carry out an advanced parametric design. However, withintraditional project delivery methods such as design-bid-build, the design may not alwaysbenefit from the constructor’s input and knowledge as suggested by the Construction Indus-try Institute and the advocates of constructability. In fact, while the architecture firm hadperformed their own explorations into the detailed design process of the panels and theirconnections that work was not transferred to the construction manager and subcontractors88due to design liability reasons. Thus, the design process past the point of architectural de-signs, lacks any parametric reasoning and is prone to human error. Based on the findingsof our productivity study in phase II of this project and through working with the panelinstallers, we have calculated an estimated total of $48,700 in potential work to be saved -by the panel installation subcontractor alone - through eliminating all delays. This amountsto a significant portion of work for the panel installation subcontractor, according to theirmanaging director and worth their effort to attempt to influence the detailed design as earlyas possible in their future projects.The use of advanced parametric design tools in the architecture industry is proving tobe a game-changer in areas such as early conceptual design of high-rises Gane and Haymaker(2007), in rapid building design prototyping (Burry, 2003), in design iterations and balancingcontextual requirements of tall buildings (Dritsas et al., 2006). While most of the work seemsto revolve around rapid generation of design solutions and easier development of designiterations, there remains other benefits to be gained from the use of parametric design toolsand their computing power. As demonstrated through this case study and the work ofan international architecture firm in a high-rise facade design, advanced parametric designtools can offer a means to achieving architectural design intents while meeting design-rulesaimed at improving constructability in design. Although less prevalent, academic researcharound the use of parametric design for constructability exists but may be too specific forgeneralization across the board (Shelden, 2002). While we also examine a specific projectand a specific type of constructability issue, the idea of associating rules and constraints thatembody engineering significance with design is widely applicable (Shah and Mntyl, 1995).893.6 ConclusionThe use of parametric design in architecture is not a new phenomenon by any means. Theextents to which users are able to harness the abilities of parametric design, however, havesignificantly grown in the past few years with the advent of tools such as Grasshopper 3D andDynamo. Using visual programming language (VPL), gives users significantly more powerand access to advanced functionalities for realizing algorithmic solutions to building designproblems. While the architecture industry has been benefiting from the use of such tools,the engineering and construction industries have been slower in adoption of this technology.Understandably so, adoption of any change to a well-established process requires the abilityto quantify the impact in order to show the value for the proposed change. In this paper,we are able to demonstrate the potential for improvement in the productivity of the panelinstallation cycle by linking the identified constructability issues with an algorithmic way ofreasoning with the problem which, in utilizing advanced parametric design tools, can resultin a design improved for constructability.In order to refine the proposed framework for utilizing computational design in im-proving building constructability, future action-research could prove useful in measuring theimpact of the proposed design solutions derived computationally. Of course, it is under-stood that the research was only carried out on one project which limits the generalizabilityof the potential for improvements in productivity. However, application of the conceptualframework developed in this study in an action-research project could yield to further con-firmation. It is our belief that the proposed framework could prove more useful for problems90of repeating nature where the power of computation can be utilized to dynamically solve,using the same set of rules, many of the same type of problem.91Chapter 4Conclusion4.1 SummaryThis research project aims to provide an understanding of how advanced parametric designcan be used to improve constructability of building designs. The project started off as a testof our industry partner, an international architecture firm’s hypothesis, that they have beenable to significantly improve constructability of building facade system by using advancedparametric design tools and translating their design into an algorithmic problem with whichthey can associate constructability improvements. Therefore, we set out to examine theconstructability of the building facade production, specifically the precast concrete facadepanels. As with most other construction projects, a direct comparison of this buildingfacade with another similar project proved to be impossible due to the uniqueness of everypotential comparison project. As a result, the planned work schedule is used as a measureof our test. This is considered to be a valid test of our hypothesis because, the planned workschedule is determined to be a significantly improved as compared to the initial facade design92which did not benefit from the use of advanced parametric design. The facade system, asit was initially designed, did not meet scheduling or budget criteria which resulted in theimprovements. Early findings of this project pointed towards a confirmation of the hypothesisunder examination. However, the panel installation observations lead to an unexpectedfinding: there were recurring challenges in the panel installation production cycle that werecausing significant delays. At this pivotal time in the research, the scope of the project wasexpanded in order to take a deeper look into the challenges.This thorough examination of the panel installation production cycle took the formof a productivity study using Adrian and Boyer (1976) Method Productivity delay Model.Through this analysis, we were able to measure the impact of the observed delay types onthe panel installation process. The delay types were then categorized into constructabilityissues versus non-design-specific delays. An analysis of the causes for each delay type iscarried out in consultation with the panel installation subcontractor and the solutions areidentified. Finally, we explore a conceptual design approach in our attempt to understandhow the use of advanced parametric design tools can improve constructability of buildingdesigns. This approach is carried out on the most impactful constructability issue identifiedand the design outcome is presented and compared to the panel manufacturer’s detaileddesign in order to demonstrate the improvements. The resulting design is also checked withthe panel installers as an expert in the panel installation process which also confirm theimprovements and potential for significant improvements in the delays as a result of thetarget constructability issue.934.2 Limitation and Future WorkThe nature of the building project under study could have allowed for an action researchproject as it consists of two almost-identical towers clad with the same facade system. Hadwe been able to perform phase II of the productivity study on the first tower, identifiedconstructability issues, and influenced the detailed design of panels with the proposed im-provements for the second tower, we would then be able to validate our conceptual approachto design using advanced parametric design tools with measured data. However, since thepanel manufacturer had the space to store manufactured panels, there were efficiencies gainedin the panel fabrication process by manufacturing all panels consecutively without any gaps.Furthermore, our conceptual design approach is an attempt at solving a specific con-structability issue on a specific building project lacking a formal structure for generaliz-ability across the board. While this case study adequately serves as an exploration intohow construction knowledge can be incorporated into an advanced parametric design sys-tem to improve constructability of building designs, this research area could benefit from aninvestigation into formalizing this approach for generalizability.94BibliographyAdrian, J. J. and Boyer, L. T. (1976). Modeling method productivity. Journal of theConstruction Division, 102(1):157–168. 13, 48, 51, 52, 68, 70, 71, 93Bolviken, T. and Koskela, L. (2016). Why Hasn?t Waste Reduction Conquered Construction?In 24th Annual Conference of the International Group for Lean Construction, pages 3–12.8, 42Bonoma, T. V. (1985). Case Research in Marketing: Opportunities, Problems, and a Process.Journal of Marketing Research, 22(2):199–208. 16Burry, M. (2003). Between intuition and process: parametric design and rapid protyping. InArchitecture in the Digital Age : Design and Manufacturing, pages 147–162. Spon Press.89Construction Industry Institute (1986). Constructability: A Primer. Publication (Universityof Texas at Austin. Construction Industry Institute). Construction Industry Institute. 4,38, 3995Cooke, T. (2013). Can knowledge sharing mitigate the effect of construction project com-plexity? Construction Innovation: Information, Process, Management, 13(1):5–9. 36,37Diaz, S. (2016). Computational Design in Construction. 19, 59Dozzi, S. P., AbouRizk, S. M., National Research Council Canada, and Institute for Researchin Construction (Canada) (1993). Productivity in construction. Institute for Research inConstruction, National Research Council, Ottawa. 51Dritsas, S., Charitou, R., and Hesselgren, L. (2006). Computational Methods on Tall Build-ings. In 24th eCAADe Conference Proceedings, pages 778–785, Volos, Greece. 89Dubois, A. and Gadde, L.-E. (2002). The construction industry as a loosely coupled system:implications for productivity and innovation. Construction Management and Economics,20(7):621–631. 37Fellows, R. and Liu, A. (2015). Research methods for construction. John Wiley and Sons,Inc, Chichester, West Sussex, United Kingdom ; Malden, MA, USA, third edition edition.16, 17Fischer, M. and Tatum, C. B. (1997). Characteristics of Design-Relevant ConstructabilityKnowledge. Journal of Construction Engineering and Management, 123(3):253–260. 4, 5,39Flyvbjerg, B. (2006). Five Misunderstandings About Case-Study Research. QualitativeInquiry, 12(2):219–245. 1796Formoso, C., Blviken, T., Rooke, J., and Koskela, L. (2015). A Conceptual Framework for thePrescriptive Causal Analysis of Construction Waste. In Seppnen, O., Gonzlez, V. A., andArroyo, P., editors, Proceedings of the 23rd Annual Conference of the International Groupfor Lean Construction, number 23 in IGLC, pages 454–461. IGLC.net, Perth, Australia.8, 42Gane, V. and Haymaker, J. (2007). Conceptual design of high-rises with parametric methods.Predicting the Future, 25th eCAADe Conference Proceedings, ISBN, pages 978–0. 6, 37,40, 54, 89Gibson, G. E., McGinnis, C. I., Flanigan, W. S., and Wood, J. E. (1996). Constructabilityin Public Sector. Journal of Construction Engineering and Management, 122(3):274–280.4, 5, 39, 40Gidado, K. I. (1996). Project complexity: The focal point of construction production plan-ning. Construction Management and Economics, 14(3):213–225. 1, 37Gransberg, D. D. and Douglas, Edward E., I. I. (2005). Implementing Project Constructabil-ity. AACE International Transactions, pages PM11–PM14. 4, 39Griffith, A. and Sidwell, A. (1997). Development of constructability concepts, principles andpractices. Engineering, Construction and Architectural Management, 4(4):295–310. 4, 5,39, 40Hanlon, E. J. and Sanvido, V. E. (1995). Constructability Information Classification Scheme.Journal of Construction Engineering and Management, 121(4):337–345. 497Harris, L. C. and Ogbonna, E. (2002). The Unintended Consequences of Culture Interven-tions: A Study of Unexpected Outcomes. British Journal of Management, 13(1):31–49.16Hartmann, T. and Fischer, M. (2007). Supporting the constructability review with 3d/4dmodels. Building Research & Information, 35(1):70–80. 4Hoffmann, C. M. and Joan-Arinyo, R. (2005). A brief on constraint solving. Computer-AidedDesign and Applications, 2(5):655–663. 8IPENZ (2008). Constructability. Technical report, The Institution of Professional EngineersNew Zealand. 38Jabi, W. (2013). Parametric design for architecture. Laurence King, London. 6, 37, 40Kim, J. and Wilemon, D. (2003). Sources and assessment of complexity in NPD projects. Rand D Management, 33(1):15–30. 1Koskela, L. (1992). Application of the new production philosophy to construction. Number 72.Stanford University Stanford, CA. 7, 8, 41, 42Lam, P. T., Wong, F. W., and Chan, A. P. (2006). Contributions of designers to improvingbuildability and constructability. Design Studies, 27(4):457–479. 4, 5, 39Larson, E. and Gobeli, D. (1989). Significance of project management structure on develop-ment success. IEEE Transactions on Engineering Management, 36(2):119–125. 1LeCuyer, A. (1997). Building Bilbao. Architectural review, 202(1210):43–45. Last updated- 2014-04-22. 3798O’Connor, J. T. (1985). Impacts of Constructability Improvement. Journal of ConstructionEngineering and Management, 111(4):404–410. 5, 38, 40Pulaski, M. H. and Horman, M. J. (2005). Continuous Value Enhancement Process. Journalof Construction Engineering and Management, 131(12):1274–1282. 4Pulaski, M. H. and Horman, M. J. (2008). Constructability Knowledge Management inSustainable Design. In Yang, J., Brandon, P. S., and Sidwell, A. C., editors, Smart &Sustainable Built Environments, pages 234–244. Blackwell Publishing Ltd, Oxford, UK. 4Radtke, M. W. and Russell, J. S. (1993). ProjectLevel Model Process for ImplementingConstructability. Journal of Construction Engineering and Management, 119(4):813–831.4, 5, 39Raviv, G., Shapira, A., and Sacks, R. (2012). Relationships between Methods for Con-structability Analysis during Design and Constructability Failures in Projects. pages 515–524. American Society of Civil Engineers. 4Shah, J. J. and Mntyl, M. (1995). Parametric and feature-based CAD/CAM: concepts,techniques, and applications. Wiley, New York. 48, 89Shelden, D. R. (2002). Digital Surface Representation and the Constructibility of GehrysArchitecture. PhD thesis, Masachusetts Institute of Technology. 89Tatikonda, M. and Rosenthal, S. (2000). Technology novelty, project complexity, and prod-uct development project execution success: a deeper look at task uncertainty in productinnovation. IEEE Transactions on Engineering Management, 47(1):74–87. 199UBC Campus and Community Planning (2010). The UBC Vancouver Campus PlanPart3Design Guidelines. 18Wong, F. W., Lam, P. T., Chan, E. H., and Shen, L. (2007). A study of measures to improveconstructability. International Journal of Quality & Reliability Management, 24(6):586–601. 4, 5, 39Woodbury, R. (2010). Elements of parametric design. Routledge, London ; New York, NY.5, 6, 8, 37, 40Yin, R. K. (2003). Case study research: design and methods. Number v. 5 in Applied socialresearch methods series. Sage Publications, Thousand Oaks, Calif, 3rd ed edition. 16, 43100

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0308669/manifest

Comment

Related Items