Making Good Assumptions For an Energy Model

Modelling an existing building has unique challenges. In some ways it can be easier than modelling a building that doesn’t exist yet because everything is fixed – there is no architect, engineer or owner to make last minute changes. Other aspects are more difficult, however, because reality is often “messy” and full of unknowns. No matter how detailed an audit is performed or how many piles of drawings are available, there will always be significant information missing. It’s not just simple quantities like the insulating value in a wall either. It can sometimes be complicated values such as the cycle times and loading of compressors, pumps or other equipment. This requires making assumptions and modelling based on these assumptions. The outputs of the model are dependent on the assumptions input, so making good assumptions can mean the difference between a model that fulfils its purpose and one that doesn’t. A model for a large building can involve dozens of assumptions – how can so many unknowns be filled in reliably?

Types of Assumptions

Before we go deeper into the question of how to make good assumptions, it would be useful to understand what types of assumptions are often required. This will vary from building to building, of course, but some types of information tend to be missing more often than others.

  • Wall Insulation/Window Properties: In many cases, drawings will not exist for a building. They may have been lost, damaged or were never passed on to a new owner. Most audits do not involve tearing open walls to determine wall construction either, so the layers of a building envelope and their respective properties are often unknown.
  • Schedules: The occupancy patterns of a building are difficult to determine and lighting use is not always easily quantifiable.
  • Equipment Runtimes: The runtimes or cycle times of equipment cannot always be determined easily.
  • Plug Loads: The plug loads or receptacle loads in a building vary wildly and are very difficult to quantify.

Reduce the Number of Assumptions Required

The first step is to reduce the number of assumptions that are actually required. This is done by ensuring as detailed an audit as time and budget allow, as well as making sure the essential documentation is collected. This can include appropriate (and up-to-date) construction drawings, detailed control sequences (including schedules) and notes from interviews with site operations personnel.

If information cannot be gathered on site, take advantage of sales people, contractors or other industry contacts. For example, the local Carrier sales rep will most likely know the types of compressors typically installed in their units, even if the nameplate has worn away. With good preparation and planning, many assumptions can be avoided altogether.

Prioritize Impact

Once you’ve eliminated as many assumptions as possible, next you should figure out which assumptions will have the most significant impact on your model. This will vary according to the purpose of your model and is difficult to generalize. If you are primarily interesting in modelling the domestic hot water systems, for example, you will not be as concerned with the insulating value of the walls. Focus your time on the assumptions that most significantly affect your application of the model.

Standards, Rules of Thumb and Past Experience

This is where the real assumptions are made. Many values have rules of thumb used in the engineering or construction world and these may be used if no other source of information is available. The better approach is to use past experience, but this requires, well… experience. If a past project had a specific piece of equipment and it had good runtime logs, you can try to judge the similarity of your current project in order to make a qualitative comparison. Then use this comparison to make an assumption about the current project.

Another source of information are various standards and other publications. For example Canada’s Office of Energy Efficiency has significant data on the typical energy use of various appliances. Product manufacturers may also publish rated energy consumption at nominal values which can be adjusted for the given conditions. Product sales reps can also be useful at this stage as they may be able to give insight on the detailed inner workings of their product.

Insights into the approach for choosing values can also be described by the following quote:

If you eliminate the impossible, whatever remains, however improbable, must be the truth. – Spock

Through a process of elimination, then, unreasonable values can be discarded leaving only a relatively small number of potential values. This quote brings up an important thing to understand about existing buildings – most buildings are broken in some way! Through a combination of errors in design, construction, maintenance and user behaviour, there will always be something that behaves differently than you think it “should”. Did the maintenance guy override the control system so that it only looks like that pump is off at night? Are occupants leaving windows open constantly? Is that outdoor air damper locked closed in an occupied building – against code? Any number of strange, unpredictable and sometimes just plain wrong situations can be found in real, existing buildings. So, after you’ve eliminated the impossible, whatever is left has to be the truth no matter how strange it may be. Experience and detailed detective work will help make this process of elimination easier.

Calibrate and Iterate

Once the model is ready, it must be calibrated/validated. This involves comparing the predicted energy use of the building to actual energy use data. If the modelled and actual energy use are not close enough (what defines close enough is a topic for another discussion) the model will need to be adjusted. The ways in which the model differs from the actual data will give clues as to how to change the model and this may indicate that one or more of your assumptions was incorrect. This process is iterative and many alternate assumptions may be required before the final numbers are accepted. This process will also reinforce some assumptions and can help to confirm that a particularly strange behaviour is occurring as your suspected.

Closing Words on Good Assumptions

One last piece of advice is to document and report all the assumptions you make. Clients will sometimes require this, but even if they don’t it’s good practice for several reasons. First, you may leave a project on the back burner for a while then wonder what your were thinking when you take a fresh look at a strange value. Also, another modeller may take over and not understand why you chose the values you did.

Your model may appear to have a thousand unknowns, but by following the basic process outlined here you should be able to eliminate many of them, then use all the resources at your disposal, and the process of elimination, to make good assumptions. The better the inputs of a model, the better the outputs and more useful the results!

This entry was posted in Energy Modelling on by .

About Matt

Matt is a mechanical/building engineer who specializes in whole-building energy modelling, energy efficiency and solar buildings. He’s worked for about 9 years analyzing the energy use patterns of buildings. He studied the energy performance of a low-energy, solar house for his Masters thesis. Matt has experience with EnergyPlus and eQuest.

Leave a Reply