DesignReForm

I guess I'll be the first to post a topic on Autodesk Inventor.

I work at a design-build firm, StructureCraft Builders. www.structurecraft.com
We design, fabricate and erect structures which are mostly composed of wood. We use Inventor for making the concept model, final model, construction documents and shop drawings. For some projects, such as the Richmond Olympic Oval for the 2010 games, I also used the model to indirectly drive the automated production of roof panels by producing and sending fabrication data from the model to PLC machinery which we had custom-designed for this project. We also use Inventor for creating material take-offs and determing information about the design, such as center of gravity, total area of material, etc.

Here's a screen capture for part of the Olympic Oval project:


Here's a render (using VIZ and VRay), of a singe roof panel, same project:


And here's a picture from earlier this year from site:


The roof panels vary in design and overall dimensions, depending on their location, for several reasons. These panels are designed acoustically, and the pattern of wood material in each panel is based on that, as well as on structural performance, and also the architectural aesthetic quality required, which involved creating an overall pillow effect, while describing the shape of feathers of a Heron, which is the official bird for the city of Richmond.

There's lots more to write about this project in particular, but the main reason why I thought it would be a neat idea to share this on this forum was because I'm convinced that other have used Inventor for these kind of projects, and I'm curious to meet them and learn about how they've used this software, or programs somewhat similar like Digital Project for example.

I have been doing a lot of research on Inventor lately and it is GREAT to see this. Do you guys use Revit also.  We use Digital Project but are very interested in the interoperability between Revit and Inventor. Was the fabricator using Inventor also?  Did you files for straight to the fabricators.

Thanks so much for posting this I'll post it on the main page.

Dave

Thanks for your comments Dave.
We happen to be the fabricators, as well as the structural designers. Our design office is at our shop.
We don't use Revit, mainly because Revit doesn't seem to be useful for creating shop drawings. It also seems that the more intelligence Autodesk tries to build into Revit, the less flexible it becomes for more custom design projects such as the ones we undertake. I'm also not crazy about the fact that a project in Revit is one single file, and that Revit's API is not very well developed yet. We may use Revit in the future for creating general layout drawings, but not for modeling or shop drawings. In order to connect a Revit model with an Inventor model, I guess a practical way to do this is by writing a simple application which takes shared parameters from Revit and creates or updates their counterparts in the Inventor model, where the Revit model would only include general geometry of the project, plus geometry by other consultants, and the Inventor model would be a highly detailed model like the ones we currently produce, that includes every little piece of wood and metal in the structure. And the connection would probably have to be 2-way to be of any use. But there is probably better of ways of doing this, I've never really used Revit before so it's hard to say.

Out of curiosity, I would have expected Digital Project to be capable of handling modeling and drawing creation, why do you also use Revit? Is there something that Digital Project lacks, or is not efficient at doing?

We did exactly what you said might work and it did.  We had simple representational models of railings in Revit and then built the more details versions with fabrication data in Inventor.  We (not me, Josh Emig and Steve Sanderson) then wrote an application to share the parameters through excel.  We basically exported a schedule from Revit and Filled a design table in Inventor.

Since it's becoming increasingly trickier for us to send data directly to fabricators we are constantly trying to figure out the best data to extract from Revit that is relevant to a fabricator. So really it becomes more about the parameters and not the geometry.

The reason we still take DP models back to Revit is for drawing production. DP has a pretty traditional drawing extraction process which we try and avoid. Revit hands down is the best for getting architectural drawings (not fabrication data) directly out of a 3d model. As an interoperability effort we are trying to get Inventor into our workflow. 

Cool. I knew that Revit was very good for creating drawings, I didn't know that there was that much difference between it and DP. I'm very interested in knowing more about DP's drawing creation process.

I totally agree about the parameters being shared as opposed to the geometry.
In Inventor, we often share parameters only between models in order to achieve very simple and light objects to deal with, since some of our projects involve thousands, even millions of parts. This is especially useful when creating what Inventor calls iParts, I think SolidWorks calls them Configurations. These are models that have a table of values, like a schedule, where each record, or row in the table, or "child" as Inventor calls it, represents a variation of the model. I used this in the wave panels for the olympic oval, where there were at least 100 parameters changing from one panel to another, and there were over 150 different types of vees, so the table was quite large. I like working this way though, it allows me to have an overview of all variations of a model, and when design changes happen, which is quite often for the kind of projects we do, then I can better understand how that change affects the various components of our structure. Often these changes originally appear as a simple dimension change, but it then propagates throughout the models in unpredicted ways, and can result in geometry errors which are not immediately obvious and needs some time to understand and solve. But then often design changes are very visible just by observing how the model changes on screen. Each situation is different, I'm sure you've had your share of changes as well.

I guess you might say that there are two general ways of sharing model data from one app to another. The first would be to create a model, say in Inventor for example, based on information from a Revit model, but which doesn't need to maintain a bi-directional relationship. Then the second would be the bi-directional connected Inventor model.
When making a bi-directional connection, you need to add alot more time making the model "robust" enough to take on changes, which means more use of theoretical geometry, such as points, planes and axis, and more thought into the possible ways this geometry could change.
That way, if the fabricator needs to make changes to the models, and have to update the architects' model, the same model can be used throughout the design process, even if radical changes to the project happen. This is common in our workflow, but not between architect models and ours, but within our models, connecting the design process of other consultants, with that of our structural engineers, and with our manufacturing department.

I assume the same scenario would happen if you had a Revit model bi-directionally connected to an Inventor model. The Inventor model would need to be prepared to take on radical design changes from the Revit model, otherwise it would have to be re-modelled. This is only a big deal if the collaboration between the architect and the fabricator happens early on in the project delivery timeline, which is our case because we are a design-build firm.
Steel fabricators, on the other hand, usually receive projects when most of the design is already finalized, and they just need to create models based on data from construction documents which are already issued for construction. If any design changes happen, the steel fabricator detailing department will probably charge for making those changes in their models and drawings, since the whole idea is that they work with final designs. In this example of steel fabricators, although not all fabricators are the same, in general they don't get involved very early in the project, and this is the same for most fabricators.
If you had a solid workflow for setting-up bi-directional Revit/Inventor models, then you could have information ready for fabricators much more ahead of time, and you'd be able to babysit the design changes as reflected in the fabrication model and drawings, plus you'd have a more detailed understanding of how these changes affect costs. So I guess it's probably well worth the effort. But it's alot of work, you want to be able to set it up so that you're not limiting your design options in any way, but at the same time not destroying your fabrication models after each radical design change.

With a bi-directional workflow like this, you should probably look into using Inventor to setup the overall geometry for projects that have very complex shapes, and let that drive your Revit model via parameters. This would allow you to play with design ideas and see immediate results, without having to deal with huge file sizes, and when you want to see this reflected in the Revit model, just use your custom app to update.

I posted a long winded message related to this topic at another part of the forum but I removed it and figured Id weave it into here.


Dave, I am unfamiliar with Inventor and was wondering how much more detail inventor was able to give you? It sounds like Revit is able to give you enough detail to document but you take it to the next level by generating fabrication data in Inventor, does this basically just give you tighter control of the final product by sharing your Inventor parameters with the fabricator?   

Santiago, thanks for all the input.  You're absolutely right about the process maybe being a bit too time consumming to be worthwhile.  It is also great to see a pure workflow like the one you have within your office.  For a bit of background, i work at SHoP as well.  Dave and I have had many discussions about the interoperability issues between these softwares, and i think the ideal situation (in a perfect world!) is that the architect's efforts in modelling can be handed to the fabricator at some point.  What i mean by this is that the modeling that we do to produce our deliverable set of drawings in revit, should be the basis for the models in inventor.  Revit bases its modelling on 'families'.  these are the objects that rule revit models.  The interoperability that would be ideal is if revit used Inventor as its family editor.  this way, the architect can put in as much or as little detail as he wants, but still be able to hand off something to the fabricator.  the fabricator then would open a native inventor file and 'read' the process the architect used to set up his model.  at this point the fabricator can choose to either start over or use this file... in any case he/she has insight into the architect's intent, which i think would be immensely valuable...

Rock on Fed!

Hi Federico,
I hope to have some time soon to look into Revit. I attended the autodesk university last year, and was present at several Revit sessions. I could see alot of power in the process of creating construction documents, and how Revit is database-driven, but I didn't see anything that showed-off any modelling abilities in Revit. As a matter of fact, having to model a family via 2D-type views seemed very cumbersome. Of course, I only had a glimpse of the program, so perhaps there are modelling capabilities in Revit which I haven't seen yet.

Inventor will never be an architectural tool, I can't imagine that ever happening, but it already serves quite well as a general-purpose precision modeling program. We produce construction documents as well as shop drawings, and although we use Inventor to create both, sometimes we end up just exporting views from Inventor to AutoCAD for to quickly annotate and finalize construction documents, because it's simply easier and faster to do this in AutoCAD, granted there are no geometry changes during the process.

In regards to conveying design intent to fabricators, you would probably have to establish a standardized language to the fabricator from day 1 in order for that to work properly, where the model information is enough and no RFIs are required. Wouldn't that be nice?
I don't recall ever seeing a standardized set of instructions on the sharing of 3D data from the architect for specific fabricators, even though it holds so much potential to streamline collaboration. You might say that the architect speaks french, and the fabricator japanese, and in each RFI the fabricator is asking "what does this mean in my language?", or "You're missing a word here", or quite often "We thought you meant this in our language" and something gets screwed up.
I don't think RFIs will ever dissappear, but they can get minimized greatly, with ideas such as the ones that you may be discussing at SHoP. If early on in each project you look into ways in which this workflow with fabricators can be improved, I bet you'll soon be onto something.

I think the process of connecting programs like Revit and Inventor today seems to be a bit new, and very R&D-ish. But I can easily see this kind of workflow being standard practice in projects in years to come. We might as well get a head start.

Dave, I saw the "Scheduling with conditional statements" tutorial.
It was cool. You can't make conditional statements in Inventor, only formulas.
In order to design using conditional statements, such as "make plate thickness this much if bearing surface exceeds that much", you either have to create your own routine with VBA in Inventor, or you have to use something like Autodesk Intent or iLogic.

Can you use conditional statements for any parameter in any family?

I can see some potential there when designing starting from Revit. This model can be connected to an Inventor model for shop drawing creation and detailed modeling, but during this process of exporting or updating parameters, a series of conditional statements can be run which ensure that certain design objectives have been met which concern both the architect and the fabricator, making this 3rd party software act not only as a filter which shares only key parameters, but actually evaluates the sharing process.
This could even be used for the reviewing of shop drawings on behalf of architects, as an initial model check which looks for certain feedback from the fabricator's model which needs to meet certain design criteria as per the architect. I'm just throwing a very incomplete thought out there.

Santiago, you're right about the modeling issues with Revit.  I find revit to be most helpful in the traditional tasks of the architect.. drawing production, quality control of documentation, and document management.  what we face is that the very concept of collaboration, though a pure necessity for the mere existence of the architect, is sometimes viewed as unnecessary on the part of some contractors or fabricators.  the fine line of where means and methods actually begin and end is a discussion that is specific to any and every occasion.  the reality is that most of the tools that architects are used to dealing with are no more than representational, and so not very helpful in the world of construction.  my aims, and the reason why i believe in a future link between revit and inventor is because it can become the first integrated architecture / fabrication tool.  i'm not preaching universal standards (i'd rather people continue to innovate), but it would teach architects a lot about what it takes to get things made... these days people seem to think that making a prototype can be analogous of to final production runs for building assemblies, i'd rather we learn and teach when it is prudent to stop and hand off the baton to the next guy...

What makes this topic really exciting for me is how it suggest the tools we use can become a catalyst for change in how things are made and its not coming from the top but from the people on the ground who are knee deep in it. When I was first introduced to revit back in the late 90's during a presentation in Boston, I was sitting in a room with a couple dozen other architects, we were all intrigued by what we were seeing but not convinced it would take off. We missed an incredible opportunity as a profession, contractors saw the potential and began using revit to build the architects cad drawings for material takeoffs, coordination, etc.... Eventually mechanical and structural subs caught on and the coordination of these disciplines was done with more accuracy by the contractor and his subs then the architect and his engineers. We took a back seat and watched it go by. Still today its done the same way, while firms worry about liabilities associated with sharing there models the construction industry continues to progress. At some point this will get ironed out and architects will play a bigger roll with more to offer during construction.

It sounds like the revit/inventor workflow has the potential to make some serious change in how project teams are put together and how nonlinear relationships can be made. Hopefully Autodesk will have more on Revit Inventor workflow at University 2009. It seems like their 2008 presentation made people curious as to what was being suggested. I think forums like this one will have allot of influence on how this all takes shape. Great stuff!   

Our Ideal situation would be an excel like interface from which any program could send or receive parameters.  This could be XML based or through web services or any other simple format.  This would require each party involved to build their own models but in the end i think that is a good thing.  They would then choose which parameters were relevant to each trade. I think this would further the understanding of the projects at had and since it is all happening digitally the risk of on site error is still greatly reduced. The idea that geometry is transferred from architect -> Fabricator is a little unrealistic because the machines, staff, software and all other factor of the process with each fabricator is going to be different.  but we we are designing a curtain wall the center center lines between panels will be relevant to everyone  involved regardless of software or any process specific issues. 

In my opinion this is where BIM should be going. It really can and should be about the preservation of pertinent information involved in the design and building process.

This makes a great point.  We've been focusing on sharing of information through the very elements that are only representational of that information... in this case Dave takes my previous comment about 'architectural tools being representational in nature' and takes it a step further in suggesting that models are in themselves irrelevant... they're only the interface through which we interact with the data that we're generating/editing.  this would render the platform irrelevant, as well as the modelling culture.  all you're saying is that the 'raw' data that everyone works off of lives in some central location.  of course you have to still define roles in terms of who creates that data and their boundaries, but we would make big strides by not having to force anyone to use any specific software...

Yup