Reference > Marc Rosenbaum's Writings

Understanding the Energy Modeling Process:
Simulation Literacy 101
from The Pittsburgh Papers (2003)

GETTING, AND VETTING, THE OUTPUT

Previous Page: MODELING INPUTS AND ASSUMPTIONS

Once the model is built, it’s time to run it and see if the results are believable. Sometimes the output or conclusions are far from physical reality. The modeler should present the output in a form that can be read and understood by the owners and architects. If the person producing the model sends a stack of printouts direct from the modeling software that exceeds the size of the Manhattan phone directory and requires the Rosetta Stone for deciphering the information, recycle the material and ask for a concise document written in English. The output report should include energy use by month and by year for heating, cooling, domestic hot water, mechanical systems, lighting, plug loads, and other sources of electrical consumption (such as elevators). The report should show heating and cooling consumption by building component, telling how much is due to walls, roofs, windows, infiltration, ventilation air, etc. This guides us to look for the areas where we can achieve the biggest savings. The report should include a table of areas for each building component (walls, roof, windows, etc.) as a quick check on the accuracy of the take-offs.

The monthly output helps us vet the validity of the model. If cooling energy rises in the winter, something’s probably out of whack. Getting component output also promotes insight. In a commercial building, for example, infiltration is unlikely to be the largest load amongst the heating components.

One of the most effective ways to vet the model is to become familiar with energy-use benchmarks for typical buildings of the same use in a similar climate. Owners of multiple buildings can build a database of energy use by building type for both the thermal and electrical components of energy use. Architects can investigate the energy use of buildings they have designed in the past (this is good practice in any case, as it serves to inform the goal-setting process in the formative stages of the project.) Databases of building energy use are available from the government (such as the U.S. Environmental Protection Agency’s Target Finder). Some good benchmarks for quickly understanding whether the model is on track include:

• Total annual energy use per square foot;

• Annual energy use per square foot for heating, cooling, and electricity;

• Cubic feet per meter of ventilation air per person of expected occupancy; and

• Square foot per ton of cooling.

If the output seems out of bounds based on past experience, a more detailed look at the model output is in order. Can a physical explanation be constructed for what seem to be anomalous results? On a cold-climate, institutional building, for example, a modeler found no benefit to wall insulation greater than R-6. I asked for, and tried to generate myself, a physical explanation for this unlikely conclusion. No one was able to explain this result. A simple hand calculation of heating energy saved by going from an R-6 wall to an R-19 wall during the building’s unoccupied hours yielded energy savings seven times what the model showed. Something here was off, and no one could explain it physically. Usually persistence will turn up the error.

Once the base building model is operational and yielding believable results, it’s time to set the items to be examined parametrically. Decisions that need to be made about the proposed building are continually informed by the powerful ability of the robust model to answer “what if” questions. Energy modeling has the potential to be highly interactive and teach all involved while funneling the owners’ resources to the places where the most effect can be made.

Next Page: ACCURACY OF THE MODEL

Page 4 of 6