What is too much model detail?

What is too much model detail?

I remember a few years ago, I was listening to an Architect speak at an industry event in Melbourne, Australia - my apologies, I wish I could remember the name! The discussion at the time was around LOD and the varying contracts that moved the goalposts around project to project. This particular Architect went against the grain (at the time) and simply did the best they could - each and every project. This meant, some of the time they greatly exceeded their clients expectations, and others, they simply met - but internally, the goalposts always remained in the same position.

This story stuck with me over the years, and is personally something I mostly agree with.

In Australia and working for an engineering consultant, we didn't start adopting Revit (Revit MEP at the time) until it was at least released in the metric system back in 2008 (release 2009 I think it was). On top of trying to learn this new tool (while still dealing with legacy projects in AutoCAD MEP) - a problem started to occur. The model(s) were getting too large for the computers to handle and the modelling/drafting time was increasing significantly - why?

In my opinion, Revit was the shiny new toy and Draftsman like myself wanted to keep pushing the limits for how "realistic" we could make the models. Of course this work ethic wasn't really shared on all sides. Bean counters wanted as little as possible, Engineers wanted everything! In addition, there wasn't any agreed reference material that defined how far one would take the modelled effort - at least until the BIMForum LOD specification (which was developed on the back of the LOD definitions coined by the American Institute of Architects) was widely accepted by the Australian industry.

This document worked (to a degree) within the consulting world - but I don't agree working as a contractor. In my opinion, the Architect I mentioned above had the right idea! Instead of trying to argue what is and is not LOD300, 400 or 500 - forget it all, level the playing field and produce the best of your ability.

Here's my reasoning;

• As your own business, you will need to develop your own family library in conjunction with was is available or you specify via your product suppliers. Make these families to look and feel like supplied product and there is most of the criteria met. Why make or use a glorified extrusion, when a properly modelled downlight and its heatsink is now accurate, recognisable and useful in clash detection and decision making?
• As-Built requirements? Unless you are planning to export to DWG, throw away your modelled efforts and just manipulate in 2D - why not achieve as much as possible during the design and/or construction process?
• The LOD system in practice, is not a replacement exercise. BIM Technicians will start with the end goal in mind - not place one family, only to replace later. Why not use a family that is actually useful to your site team?
• Ignore contractual obligations and look to meet and exceed your own site/construction teams needs. Treat your site team as your clients and meet your "real" clients expectations as a by-product. Isn't this the BIM sale pitch in essence? Resolving potential issues in the virtual construct first?
• Take advantage of the better visualisation. When the virtual product clearly matches the proposed supplied version - clients have a much better understanding of what they are going to get at the end.
• Achieve more BIM uses. Trimble setout stations, Point Cloud BIM validation, quantity takeoff, accurate progress tracking etc are more easily achieved with a higher detailed model - otherwise it's "rubbish in, rubbish out".
• Computing power has improved out of site and your delivery team members will be further developing their model(s) whether you do so or not. You are going to need that extra RAM, Graphics Card, CPU and storage space regardless. The days of trying to limit services models to 250mb are well behind us.
• Arguing costs time, and time costs money. If your modelling team is aware what your consistent deliverable is, they can proceed - business as usual. Get those goal posts in concrete and enjoy one less argument!

That said, I still would keep one eye on practicality and use parts of the LOD system that work for you. This is best achieved by remembering it's an element based system, not a representation of the entire model. What practical purpose does accurately modelling an electric strike achieve? None. Use an annotation symbol and add what ever data you need (achieve LOD100) and save the modeling efforts for elements that count - Level of Information (LOI) is a story for another time.

At Stowe Australia we have wrote an evolving Revit family brief that is given to our suppliers and/or our specialist content creators. This helps us grow our family libraries with curated content, achieve our site teams needs and expectations and finally , in most cases, exceed our clients BIM deliverable requirements.