The second post in the series. This time I want to touch on the geometry detail level of Revit projects or as commonly known as LOD or LOG. I will leave the data aspect or LOI for another post.
I did mention briefly in the first post about one pilot project. A steel pedestrian self-anchored suspension bridge, not of conventional type but augmented to use CHS elements instead of cables, giving these “cables” some flexural rigidity. One of the issues, when we started that pilot project, was to make sure that we can deliver proper shop drawings from it and accurate schedules. However, this was in relation to the existing workflows based on CAD. Meaning that the output had to be of equal or greater quality. This naturally meant that the requirements for the detail level of the BIM model were very high. There were no BIM regulations in place nationally and while we knew about Level of Development guidelines (by BIM Forum), there weren’t any such requirements imposed on us. That was both a good thing and a bad thing, which I’ll explain throughout this post.
The project was supposed to serve as a future workflow to transition the bridge department from CAD to BIM. Hence, we had to account for possible resistance from people. The idea at the time was that if the resulting BIM model would deliver drawings of the same quality and detail level as with CAD all while replacing the need to manually estimate quantities (reducing hours on such tasks), then it should be easy to get people to transition. In terms of clients and contractors there was another potential issue. They are used to receiving drawings, schedules of certain quality, with a certain amount of details, etc. It would have been very difficult to explain at the time what BIM LOD is and why they do not need the level of detailing they were receiving from projects. So, our requirements can be summed up as – “do it better”.
High LOD level workflows are not optimized for efficiency
We did deliver the BIM pilot project with all quantities retrieved from Revit. The LOD of the geometry was excessively high. We had produced drawings from the model with close to no 2D Revit details. Sounds like it was a success (it was) but it wasn’t the most efficient way to approach BIM and productivity increases were questionable. Time saved from quantities was mostly wasted on modelling the highly detailed geometry.
Due to all the detailed geometry, our final Revit model as not performing optimally. There was way too much detail on most of our elements, including nuts and bolts, hangers, etc. We even had all the gaskets modelled for the PVC rainwater drainage pipes and modelled the stainless-steel tactile indicators for the disabled with all the grooves, so they show up correctly on drawings. Needless to say, it was excessive for a pilot project and would still be excessive for most projects today. The Revit model became heavy and loaded slowly. There were a few thousand of the mentioned elements. Just managing the data of these elements was taking too much time. Most of this detail was not technically necessary for any use case at the time. Volume/weight is not important for PVC pipes, disabled tactile indicators, and other similar items. Their quantities go by units or meters most of the times, hence these objects can easily be represented with minimal geometry.
It does get a bit more complicated for steel elements, as they usually go by weight. Let’s consider bolts, studs, etc. If you need to specify the schedules by weight, then it is best to approach this through custom parameters in the case of Revit, as deriving the weight from the volume requires perfectly accurate bolt and stud models. This is completely unrealistic. Best case would be to go by giving the amount in individual pieces. But that might depend on the constraints and traditions in your market. The beautiful part about BIM is that it also has a data layer, allowing the information to be specified without having to solely rely on over detailing the geometry.
A use case where you might want to have as much detailed BIM model geometry as possible is cost estimation. It is hard to pinpoint which part of the model should be as detailed as possible to increase the accuracy of cost estimation since projects are all unique. Certain structures such as steel based ones have significant costs related to the material and the welding, assembly works. Connections, welds, can increase the total sum quite rapidly. It just so happens that in pretty much all cases you will want to have these elements as detailed as possible as early as possible in the design stage. That does conflict with the reality of most projects though. Should be obvious, but to illustrate the point, say our cost estimation is off by just 1% from a 10mil EUR total, that would translate to 100k EUR. That is not an insignificant sum of money.
Choosing the right Level of Detail (or Development) for your BIM elements can be a balancing act between the cost estimation and all the other use cases. Nowadays, most of this is agreed upon relatively early in the design life cycle through EIR and BEP documents in the countries that already mandate them. BIM Forum guidelines for LOD are often used as a certain benchmark. They are not perfect. Therefore, as a BIM manager/coordinator always interrogate the LOD requirements in the BEP document, if there are unclear areas as to why certain elements must be modelled with such a high level of detail, ask for the use cases to be defined more clearly and use them to argue against illogically high LODs. While difficult, the same reasoning of use cases goes for managing the client who, as I hear from colleagues in other markets, often abuse the LOD 400 and even LOD500 levels. Even an As-Built BIM model does not automatically mean that everything should be represented at the true to life level.
Interrogate the use cases
My current opinion is that intended use cases are what matter most. The BIM manager or coordinator should pay extra attention to this when dealing with the parties involved. I tried to summarize a list of questions to help guide the thinking process for what LOD levels are essential:
- How many objects will be present in the Revit model or any other software? Each software is unique and scales differently in terms of performance related to object counts. Too many elements can have a negative impact on managing the BIM model, increase coordination times, increase the likelihood of mistakes, affect the performance of the model etc. If the project necessitates large numbers of objects, can their geometry be simplified?
- How much and what data should be extracted from the BIM model or the object in question? What does that object represent? Identify early on what are the key data prerequisites and model only the geometry to the level where that information can be obtained, retrieved, or derived. If you need dimensions for clearance then focus on that.
- What items will be scheduled in the model? Identify the objects that will not be scheduled in terms of weight, volume, etc. Objects that will end up in schedules as a number of units or other simplified metrics, such as length (e.g. various bolts, rebar continuity strips, etc.)
- How significant is an accurate representation of the object in question for the estimation of costs? This can define how much detail you should model within your model. If you are a contractor preparing a bid, you want to be as confident with your costs upfront as you can be. Same goes for the client if say your national road administration is that client and wants to procure some tunnels, bridges. Planning for accuracy early in the design lifecycle can pay for itself.
- Will the BIM models be used for visualizations (architectural renderings)? And will having detailed geometry in the BIM model help with this use case? Is this a large part of the workflow? It might make sense to increase the detail level of the geometry even if other use cases do not necessitate it.
- Will the BIM model be used for fabrication? Can data and geometry be extracted from it to be suitable as input for various machines? Likely, the process will still require manual work from people at the fabricator to translate the geometry to be understood by their machines. Shop drawing level models are not the same as fabrication level BIM models. The whole world is talking about BIM collaboration and integration, but the mainstream of the industry is still not there yet.
- Lastly, think about the uses cases that will become impossible with the established geometry LOD level. Will it cause any problems in later life cycle stages? Is the client, designer, contractor and owner fully aware of this? These aspects must be communicated early on.
Summary
In retrospect, I still think going for the excessively high detail level was a good choice in the pilot project. Transitioning from CAD to BIM is more about getting people to accept change as with any new technology, workflow, etc., than just the technical aspects that are easier to control, measure and evaluate. However, that was then and now it is different. General BIM acceptance has increased significantly and typically the parties involved in a project have at least some understanding of BIM. If I had to start a pilot project now, I would optimize the geometry and prioritize workflow efficiency, productivity improvements and spend more time communicating why apples to apples comparison is not possible with CAD workflows.
The LOD of your BIM model should Ideally be related to your use cases. It is not a set up once and go thing. The requirements vary with each project and so should your approach to LOD. Again, to emphasise, the geometry must serve an objective, it should be more than to just look great, be representative (though that can be a use case). Ask yourself the questions listed above and add some of your own. Generally, what one will find is that the BIM geometry detail level can be tuned down once you talk to all the parties.