A WUFI Roadmap
When it comes to WUFI, you might start out being impressed or concerned – but as we discussed at a recent LAB Event, you should probably be both.
Chris Schumacher, who led this event, pointed out that “by definition, a model is an approximation of reality”. Many aspects of reality must be simplified to make a given problem manageable and to handle uncertainty.
That doesn’t mean that simulations can’t be useful. A well-built model can be an excellent tool for a variety of purposes, including education, product development, forensics, and design (most commonly to compare the performance of different design options). WUFI in particular is a popular program for good reason: it is well-validated, has a user-friendly interface, and handles driving rain. It has an extensive database of material properties to get you started. However, the various components of a WUFI model have different limitations and degrees of precision.
To work successfully within these limitations, it is important to be clear about why you are developing your model and what information you can get from it. In general, modeling should be an iterative process:
- Identify a need.
- Define the problem in terms that can be modeled, and build a model based on that definition.
- Run your simulation (i.e. use WUFI or other software to see what the model predicts over time).
- Interpret the results and adjust the model (and possibly the problem definition, for example making it more specific).
- Run again, interpret again, adjust again, until you have enough information to proceed.
The model itself is made up of several components For newer users, both the sophistication and the limitations of WUFI make it easy to create misleading results.
Chris described some common problem areas he has seen, such as:
- Substituting materials where they are missing from the database (e.g. fibre cement sheathing board instead of fibre cement siding)
- Selecting from multiple material options without knowing what the differences are (e.g. the difference between “air space 10 mm” and “air space 10 mm without additional moisture capacity”, or “plywood USA” and “Plywood low”)
- Using defaults for surface transfer coefficients, climate parameters, calculation period, numerics
- Using a number because someone tells you it is the standard one (e.g. always setting the interior surface permeance to 10).
Expertise is critical to building useful models — in other words, models that are close enough to reality to meet the identified need. To select reasonable inputs, recognize when a model is “off’, and make intelligent, informed choices about how to fix it, you need to understand the physics of what you are modeling as well as how the program works. It also helps to have enough field experience to have a sense of real-world materials and performance.
For newer users, Chris offered some tips to build skills (and better models):
- Assess sensitivity to different parameters – how much does the result change if you change a certain parameter? Does a giant change in the parameter have zero impact? If so, that parameter is likely not what’s making things go wonky.
- For material properties, it can be helpful to contact manufacturers. Front-line salespeople may not be able to tell you anything that isn’t listed on the package, but many manufacturers have technical experts available who can answer more detailed questions.
- It helps to begin with known information. Start by modeling a system with known performance characteristics. Try to resolve any discrepancies to find possible problem areas in your model.
In the end, though, Chris emphasized that even the best-designed model cannot, by itself, provide a good answer to most questions. Rather, the analyst must determine the best solution for the specific need, using WUFI as one tool along with theoretical knowledge, past experience, etc.
The post A WUFI Roadmap appeared first on RDH Building Science Laboratories.