Cost Analysis

It's taken a while to get this up, but I've finally finished doing the cost analysis for the various models that were explored in the previous assignments.

In order to accomplish this I created a database of assembly costs from a current RS Means catalogue, and then used Revit to calculate the amount of each assembly in each model.   Once the basic system was set up it was rather quick to generate the data for the various iterations.

For the real estate data I did a sampling of property for sale as listed on the online MLA database, and then calculated an average for each of the basic categories of land that I identified.

One thing that this analysis is missing is a calculation of the mechanical systems for each option.  This is especially important because in the last analysis it seemed that the system selection had the greatest effect on energetic performance.  Another thing that is currently missing is a synthesis of the analysis from the two assignments:  energy vs. cost.

The Graphs

The first set of graphs illustrates the relative assembly costs of the various building forms.  Of note is that in the stacked model the lifts and extra floors become a sizable percentage of the cost.

The second set of graphs represent the cost to upgrade each building type with 4" of eps insulation over all exterior walls and ceilings, as proposed in the "super insulation" option.  It also shows the costs of the options relative to each other.  The stacked option is higher than the others, but not exorbitantly so.  The Second graph in this series shows the cost of each assembly in relation to each of the other options.

The third set of graphs looks at the cost of construction in relation to land.  Considering only the land needed to build the store (as parking would be consistent across the models)  the stacked option begins to look much more attractive in all cases except where land is dirt cheap.

Timelines

So it looks like I am running about two weeks behind in the timeline that I outlined earlier.  This is not really that bad, though, considering that the timeline was really quite ambitious to start with.

So, I think that what I will do is eliminate the optimisation exercise from the course.  That will give me two weeks to do the economic simulation (which seems feasible) and have me finished the reporting work by Nov. 25.  That would then give me 3 weeks to work on the final essay and reflection.

The optimisation can be done graphically and somewhat informally using the graphs I have prepared, so it won't really need to disappear.  It will just be simpler.

Energy analysis

I've used a combination of Revit and Green Building Studio to quickly mock up some energy studies for four massing strategies for the fictitious big box store.

One thing that this involved was rebuilding slightly simpler versions of the models.  After some initial efforts with Ecotect, I found that the studies weren't really concerned with where structure, curtain wall doors or furniture might be so I removed that.  This made it easier to manipulate the models, and also gave the degree of resolution that the GBS seemed to be operating at:  rough, overall initial calculations.

I also resized the building to be 5000m2, to bring it in line with average sizes for North American big box stores.

GBS lets you do different "design options" for each model to aid in quick comparison.  I played around with different wall insulation systems, better HVAC systems and rotated the building on the site in the various options to get an idea of where the savings might lie.  It seems that the stacked, two floor layout with a much higher wall insulation level and a high-efficiency heating and cooling system could provide annual energy savings in the order of $30,000 versus the usual flat and wide building type.  With a 50 year life-cycle that would translate into a $1.5 million dollar savings.

My use of the tools are rather clumsy at this point, so I don't really know much about the particular heating system that I specified other than that it has high efficiency, but they have some rather detailed specifications.     It would also be possible to add in site massing of the surrounding buildings, to see how that would impact shading.

GBS works on a cloud system that compiles weather and local utility data to input into the model analysis.  It is a simple process where you basically just upload your model and then their systems handle the computation and storage of the data, which you can then view and download in various forms.  It is a subscription based service.  I downloaded a spreadsheet of key performance data for my various options, cleaned it up in excel, and then used that data to generate the graphs above in illustrator.

Interestingly, the new cloud paradigm seems a double edged sword for users.  On the one hand it stores and manipulates your data off-site, so your system does not get bogged down in the processing.  It is also a means for data to be constantly updated.  On the downside, one needs constant network connectivity, there is an ongoing cost, and it is harder to "hack" (ie. get a free version of the program) than the current standard of on-board software.

As for a BIM feature, it could be nice to have analysis done as a pay per use service, so that one does not need to invest several thousand dollars in a piece of software that might be used for only a couple projects a year.  There also seems to be alot of effort that has gone into making the process relatively seamless and quick, so that analysis can be done on the fly during a project and does not necessarily have to be a central concern / team capability.

GBS Energy Analysis, First Run

I managed to get GBS set up today.  Here are the results of the first run of the project through the program:

  As easy to use as the system is, my skills with the modelling are still relatively rudimentary at this point, so take these results with a grain of salt.  

There is a sizable cost for energy use throughout the year, which seems to be an area where more efficient design could bring cost benefits.  As well most of these costs are for heating and ventilation.  The construction of the project was modeled as concrete with rather minimal insulation.  Some changes to these assemblies might produce better results.  

Energy Analysis: Choosing a Platform

Today I have started on the energy analysis portion of the reporting project.  The first step has been selecting an appropriate software / means to do so.  What I have discovered in the Autodesk lineup (which seems a logical choice for ease of interoperability) are the conceptual analysis tools in Revit, Ecotect, and the Green Building Studio.

Ecotect is the software that I chose, given that it is considerably more powerful and will produce the most detailed reports in the end.    It also offers a way to interface with the green building studio, so it's a sort of two for one deal.

Both the tools built in to Revit and the Green Building Studio are network based services running as a "cloud" process.  Basically this means that the necessary number crunching and updated databases of local energy costs, weather etc. are done off of the computer used to design the project.  The services are both subscription based, meaning that they will be an ongoing cost for the architectural operator, but also that the databases of relevant information will be continually updated.

The GBS will take the project at various stages of design and return graphic analysis of various energy related performance parameters-  estimated energy use, cost of that in relation to local rates, potential for energy generation, water use and cost.

The Autodesk Files

I found a rather interesting bit of Autodesk history that is not entirely BIM related, but presents a rather interesting historical snapshot of Autodesk as a start-up company:  the Autodesk Files.

http://www.fourmilab.ch/autofile/

Besides annotated company memorandums, business plans and transcriptions of speeches at internal meetings, there's some images of trade show booths in the mid 80s and even images of their migrating headquarters as they grew.

Big Box Progress

Things are starting to move on the big box store.  Today I got the space planning done and the basics of the model together.  I had some fun getting things parametrically constrained so that as I moved walls and grid lines around, decisions that I had made regarding positioning would be retained.  I haven't figured all the best ways to do things, but its starting.

My next step will be to assign the rooms and finishes, take a look at assemblies and structure and then I should be able to start extracting performance data.

Right now I have 12m x 10 m structural bays, a receiving / storage area in the bottom left surrounded by offices and bathrooms, and an entrance at the top with checkouts to one side.  I still need to add a big garage door in the receiving area, probably some lights in the sales floor, and some signage / canopy at the front.

Block / Family Libraries

Revit is great and all, but what it really needs is a big online repository of families, so that you can just drag and drop them into your project instead of sourcing from all kinds of manufacturers.

Oh, wait... there already is one?  And it's good for sketchup, AutoCAD, 3DSmax as well?

Thank you very much, Autodesk.

http://seek.autodesk.com

Also a good one, with more user submitted content:

http://www.revitcity.com

Theorising Object Oriented CAD / Analytic Worldview

You wouldn't think that 18th century religious architecture could help understand what BIM is all about, but it turns out that it can.

In his essay "The Freestanding Column and 18th century Religious Architecture"  Antoine Picon takes the position that certain architectural modes of design are possible only within certain worldviews.  He establishes a history of the column type that appeared in France that combined the post and beam articulation of Greek temples with the slenderness of Gothic cathedral structure.  The column's articulation was that of a series of objects - ie. a post that held up a beam - rather than a  monolithic, synthetic / solid articulation like the Romanesque style, and that in order to achieve the unprecedented scale and slenderness of the columns (which was made even more difficult now that the beams were of stone instead of wood) required engineered iron and bars be inserted into the construction.

The articulation of the pieces, Picon says, was related to new modes of analytic thought.  Engineering and Architecture at that time were changing from a practice of tradition (based on Classical / Vitruvian models of scale and proportion that made a building work because it was within the ballpark of models that had been refined through trial and error) to one of experiment driven by analysis.  Instead of thinking of force flowing as a sort of liquid down piers and walls as had been customary, force had become quantified and, and the building was seen a series of parts that transferred a certain amount of force from one piece to the next.  Engineering had become more "scientific" in its practice, and now used experimental methods such as adding rebar to transfer loads to the ground.



What does this have to do with BIM?  As opposed to previous generations of 3D modelling programs that I have used, Revit assigns new attributes to geometric objects other than its aesthetic properties.  It is an analytic tool, looking to quantify the ways in which a building is performing.  It exemplifies a certain worldview - one both experimental and analytic - that by modelling a series of pieces that are given performance attributes one can understand how the finished project will respond.

It's not really groundbreaking to say that a medieval mindset couldn't have come up with BIM, but I think that it is important to note that the particular way in which it works reflects a social reality.  Broadly within society we value performance, predictability and experimentation, and have historically - as well as currently - used analysis and modelling in their pursui.  Interestingly, these values are exactly those typically touted as the strengths of the BIM system, whether that be within the spheres of environmentalism (through energy efficient design), real estate (via cost and aesthetic simulations, and collision and sequencing analysis) or through the execution of complex "high art" architectures (as a tool for both creating and managing complexity).

At the very least BIM and its meteorically fast adoption across the industry points towards a cultural preference for  performance, predictability and experimentation, but it can also be claimed that it will propogate them.  Through a system of micro-efficiencies, Revit encourages designers to consider performance and constructivity issues during the design phases.  By being good at and simplifying the collection of performance data, it encourages design that emphasizes these characteristics.  It doesn't make it impossible to design without regarding them;  but it doesn't have to.  Life and evolution work through the creation of tendencies by encouragement and small advantages.  It is the equivalent of a government encouraging a certain kind of investment or business patterns by offering tax breaks;  the market follows the small advantages.

It's not necessarily bad that architects will focus more on these issues;  in fact it is something that I think is very exciting and a welcome development.  But one can never do everything at the same time, so some things will be left out in place.

As well, these values all have political alliances that I will take a look at in a future post.

Object Oriented CAD p.II

Now that I have some experience with the software under my belt it is time to start asking questions of it.

What might the consequences of this object orientation be? What sorts of biases might be encoded in this way of working?

Already I have stated, Revit and BIM are production oriented softwares.  The object oriented method is definitely more suited towards rather standardised methods of construction, and so it propogates industry standard techniques.

Economic Repercussions:  Global Mindset

The online block repository would privilege those manufacturers and providers who can both produce their own blocks - the process is likely time and training intensive - and get them on the server.  Likely the server is for companies with national or international reach and so that there would be a tendency for  mom and pop operations would start to get eliminated from the project.

On the other hand, as many of the techniques in the industry are standardised, standard blocks actually might not hurt them once it comes time to tender a project as everyone does windows essentially the same way.

Earlier (but Less Decisive) Decisions:  Categories = Politics

Also encoded in the method is a sort of definitiveness to the design.  In a standard design process, the space between walls or floors can be left blank until it is figured out later on exactly what the composition might be.  Many of these questions are asked earlier in the project with Revit.  It is, of course, possible to use generic walls and floor types, but there is an increased impetus for these decisions to be made earlier.

Luckily, these decisions are not definitive.  One other consequence of the object type method is that it is very easy to go back and change decisions later on, because by changing an object type all its instances can be updated throughout the project.  Want to use bamboo flooring instead of hardwood now?  No problem.

There is an aspect of conservatism to this method, still.  Even in the way that the program asks you to decide  whether something is a wall or a floor as soon as it is drawn, the program consistently reinforces typical elemental categories and methods throughout the process.  Personally, I've had too much of the uncertainty cherished by deconstructive thinkers and don't really think that this type of politicisation of language is useful, but there are many who do.  This aspect of the program would probably be a big problem for them.

Materialistic Bias:  You are What you Eat

In Revit (at least in the form it comes "out of the box") the criteria that can be attached to an object are primarily performative:  cost, weight, fire rating etc..  If these are the criteria consistently placed in front of a designer, one can imagine that focus will increasingly be on these parameters.  It's a case of daily culture for, as the saying goes, "You are what you eat".  Somehow tweaking the available parameters to assess the qualitative goals of the client or designer seems to be currently difficult to do.

In my opinion this materialistic bias is not a bad one for architects.  While performance based architecture is slowly becoming the philosophy du jour in many academic circles, it is often underemphasized by many in the profession (academy included) and its media.  Architects are often seen as stylists;  specialists in aesthetic, meaning and the orchestration of feeling in a building project.  Performance is the language spoken by many of the other key stakeholders in the construction project, and so it is to our advantage to become well versed in both its language and delivery.

Anti-Monolithic Design / At Peace with Pieces

By forcing design to be assembly specific from the outset, the program moves away from the sort of monolithic design that could be considered a consequence of typical CAD / 3D modelling programs.  Everything is a material assembly now instead of some white plastic shape.  This could potentially translate into an aesthetic of pieces instead of plastic, monolithic form.  While OOCAD has been closely aligned with many of the expressionist projects of the late 20th c. (Gehry used CATIA, another solid modeller to produce their deformed shapes), if one takes a Modernist approach to interpreting the tool / medium this is not necessarily it's "essential" use.  The model is built up of a series of pieces or parts, not deformed from a single lump of clay.  It's "essence" is an assemblage of multiple pieces.  Thus one could imagine an architecture of multiple different expressed pieces being its product.

Object Oriented CAD

One of the defining features of Revit (and I would assume many other BIM softwares) is that it constructs the project out of a series of individually considered and attributed objects, rather than as a series of simple geometric lines and planes.  For those familiar with AutoCAD blocks, it is as if the entire project were created entirely of dynamic blocks.

What's so special about that?

It the object oriented approach, the objects are themselves subsets of data, or information containers.  In  OOCAD one can imagine the object properties available in an AutoCAD object such as a line containing more information than its geometric and formatting options.  What is both important and powerful about this approach is that it becomes possible to register and eventually understand a project as more than its geometry.  As well, changes to the types can be used to update large numbers of object instances throughout the project quickly and accurately.

A window in the project is created from the "window-variable-standard" basic object type (as seen on the top right) scaled to fit its opening and located in the project through the properties in the right panel.  The object types properties (center) can be used to modify the type's relative dimensions (in this case frame depth, window location etc.) as well as model, manufacturer and cost information.

As well, performance information about individual materials and assemblies can be registered within the 3 dimensional model.  The model can be useful to multiple members of the project team,

Properties for a wall type include: (middle) the various layers of an assembly, their function within the assembly and the way in which they turn corners;  each of those layers is also given its own material properties (left) including a graphic appearance and engineering details.

Programming specialists will likely point out that most CAD platforms already are object oriented in their programming construction, but what I mean by this is an approach to constructing building elements out of information rich and multiple smaller pieces, which already are in tune with the expected industry product types and standards, that are then transformed to create the individual building.  This amounts in my mind to an actual break from past CAD methods which were largely drafting table operations performed on a piece of software. On a drafting table, it is not possible to create the types of information rich objects that can be done with OOCAD;  this was done instead through the system of typical project documentation of drawings, annotation and specifications.

Revit First Tutorials: Complete

Here's the fruits of the first few weeks of Revit tutorials.  I've been working through the series linked in the previous post, the Autodesk Student Expert tutorials.  It's been a good crash course as to what is possible with this program and has me feeling pretty confident about starting my own project.

I definitely need to tip my hat to the team over at Autodesk Student Experts for putting together these vary informative courses.  They've definitely kick-started my studies of what can be done with this tool, and so far I've only worked through a third of their offerings.

So far I've learned to create an orthagonal model, insert doors, windows, stairs, curtain walls, create grids, change materials, create new sheets and schedules, cameras, edit wall types... really all the basics.  I am really impressed with the program at this point, and will start posting some ruminations on the its setup and prejudices over the next while as I move forward with the first assignments.

A New BIMsness Model?

According to David Celento in his Article "Innovate or Perish" (Harvard Design Review 27), in a business as usual working relationship, BIM could actually be harmful for architects.

In the USA at least, big customers like the General Services Administration (equivalent to Canada's Public Works and Government Services) have become wise to the increases in predictability and error detection inherent in a BIM process and are now requiring at least a partial BIM model for all Federally funded buildings. "Their experience is a reduction in change orders saving the GSA up to 10% of total construction costs."p.4  So for a 10 million dollar building that is an average of 1 million dollars of savings.

While it is great that this new technology is being embraced and seen as a source of value by owners and building managers, they don't seem willing to pay extra for it, and it really does take a lot more work to produce a detailed BIM model.  As one high ranking civil servant put it:   “Architects are paid to provide buildings without errors, why should they be paid more to do this?”

Umm, maybe because our market rate fees are based on a common practice model that does not involve this extra work?

But more work and less profit are only two of the three dangers that Celento warns may come with BIM.  Because sophisticated clients are now expecting increased error reduction, when errors due occur they are even more flabbergasted and prone to placing liability on the architect.  It could make architects even more on the line for errors, omissions, and general construction site chaos.

It's not all grim though.

Celento points to the potential inherent in owning the rights to BIM models.  While in the GSA projects the client demands exclusive ownership of the final model (which I assume comes from a security paranoia perspective) if the architect were to retain ownership of the design (which I understand we do in Canada) and the contract documents (which would be the BIM file, and which are also owned by us in Canada) then one would have a half finished project ready to be deployed for another client.  If BIMming requires way more front end work, then we may as well reuse as much of that work as possible.  As well, with the parametric setup of many BIM programs, tweaking of room sizes etc. could be a fairly easy process.

I could see this being especially useful or valuable with custom homes.  With a series of base models to pick from, customers could more or less take an off the rack design, adjust it to fit their space needs, and then parametrics could be used to fit it to the site.  The construction documents could be produced more or less automatically and then bam, you have a semi-bespoke home ready for construction.  The product would be less custom than a fully tailored suit, but way more unique than standardised tract homes.  It would be the design equivalent of a Freitag bag.

link to the article (highlights are not mine):  http://workgroups.clemson.edu/AAH0503_ANIMATED_ARCH/M.Arch%20Studio%20Documents/STUDIO_Innovate%20or%20Perish.pdf

A roughly drawn site plan for use in massing.  Long rectangle is the sales floor, the medium sized one is back room storage, and the smaller ones are entry, bathrooms and office.

Playing around with the massing from the site plan.

My initial observations of the program's prejudices are that there is a tendency for things to automatically snap together, reinforcing ideas of contiguous objects.  With curved objects there seems to be less control over the creation parameters than with Rhino.  Curved object creation seems to be kind of on "autopilot".  This probably has to do with the process of continually updating geometry rather than from creating a single perfect geometry.

Day One Tutorials

First day today!

Step one to learning this program will be to block things in with the conceptual massing tools.  I`ll be going over some online tutorials to do this:

general use of the tool:
via a series of tutorials provided by Autodesk on massing and rendering
http://www.youtube.com/watch?v=OutdBBIowZA&feature=related
http://students.autodesk.com/?nd=revit2011_english

lofting capabilities for curved surfaces:
http://designreform.net/2009/03/revit-2010-new-features-conceptual-mass-make-form-lofting/

and a whole bunch more tutorials at the design reform site.  I like the look of their page, and will be looking back there for future assignments.
http://designreform.net/category/tutorials/revit-tutorials/