Three levels of BIM quality checks

I like to ask people in the industry what "quality" in buildings means, and often I get an answer like this:

Quality is when a building has X. You can fill X with whatever you want. e.g. X = no skirting boards

But I think the following definition is in the context of BIM more useful:

Quality is when X fulfills the requirements of the customer(s).?

Now you can fill X with building, product,?or whatever you want, and the customer can decide what quality means.?

This definition is hard work; most customers need support articulating what quality means for them. Or only know it when they see it. Therefore the exploration phase, the concept design is so important. That's why I say that the planner has three jobs from the point of view of a client:

1. Developing good concepts.
2. Enabling client decisions.
3. Serving the construction site.

And these jobs are contradictory. For one and two, we need to explore. We need flexibility and iteration. For number three, iterations and explorations cause problems and wasted money. A conflict of interest we need to solve with proper project management.

There are mostly professional clients who build more often and know what quality means for them. Usually, they have different requirements for the developing time and after the handover. They had bad experiences and want to ensure quality during the design and construction. For this, different strategies exist:

1. Hiring experts to support the client's side. I know some organizations where more client reps than planners work ;-)
2. Formulating extensive requirement documents. Not uncommon to have 3000 or more pages to ensure everything is written somewhere and the organization is safe in the case of a lawsuit.
3. Checking requirements (semi) automated against the actual design.

This automatic check is one of the three reasons why owners ask for BIM. Using technology to ensure quality during design and construction - a risk management method for clients.

E.g., There is the risk that the business case won't work anymore when the bathrooms get too small. There is a high probability that this reduction will happen during the planning process because the wall lining thickness was underestimated and needed to change.

Knowing about this risk, we can set up different risk management strategies on different levels, E.g.

1. Selecting qualified planners?
2. Make sure that the HVAC planner formulates the requirements early on and that the architect plans with the correct sizes.
3. Formulating requirements for automatic checks. E.g., To ask the planner team to model the wall lining in a certain way so that it can be checked automatically. I would ask them to treat the wall lining like a shaft with a own IfcSpace element that can be easily checked.
4. Check directly when the space size changes. E.g., By automatically checking bathroom size against the required minimal bathroom size.
5. Set up a process to check regularly and often so that you discover discrepancies as quickly as possible and can act.

Strategy number one and number two are the usual ones. In times when we need more people working, of more and more people retiring, and of more and more people not interested in these dull quality assurance processes, it's time to start implementing strategies three to five. Or with other words, I like to call it employing the algorithm. Once the process is set up, the employee algorithm does not care how often the job is repeated!

From a company's point of view, setting up these rulesets is a way of capturing know-how for the benefit of the project and the other employees. That's where the beauty is.?

To capture this know-how, it's not enough to check for data quality. Let me give an example.?

1. First level data requirement is: An IfcSpaces with a PredefinedType of GFA, PARKING or SPACE. This would enable us to query the model to show the gross floor area, the Net floor area, and the parking area. In analogy to programming, this is a unit test.
2. Second level requirement is: To specify how the spaces need to be modeled so that the checks work. E.g., to count the amount of available parking (another project risk if we suddenly don't have enough), every parking space needs to be modeled as a single IfcSpace element. If this is the case, we can set up a check to count the parking spaces, to check if there is enough clearance (height and width) and if the road in front of the parking space is wide enough. In analogy to programming, this could be called an integration test.
3. Third-level requirement: This would enable more complex simulation types of queries. E.g., are there enough parking spaces in the whole development? For this, we need the requirements from the first and second levels as well as a requirement allowing us to count the number of flats in the development dynamically and the formula with the regulatory rules. The data requirement could be an attribute/system/zone connecting the different spaces to one flat. With this, we could query the number of flats, and the rest would be just some basic formulas. In analogy to programming, this could be called a systems test.

Currently, we mainly talk in projects with BIM about the first level. The actual value comes with two and three. Do you have good examples to share for clients' quality checks?