Skip to content NREL Buildings Research National Renewable Energy Laboratory (NREL)
Field Test Best Practices: A Resource for Practical Residential Building Science

Main menu

Role of Building Simulation in Field Testing

Preliminary Simulation Study

Use a building energy simulation program to refine initial estimates of project benefits. The researcher should make his/her best effort to accurately simulate the behavior of the ECM or phenomenon being investigated. This may involve reviewing the source code to understand how the physics is currently modeled and what the relevant underlying assumptions are. A literature search should help guide the development of a reasonable mathematical model based on available information. This preliminary simulation process is vital because it reveals what measurements are required in the experiment in order to verify and/or fine-tune the model. If there is no reasonable model to use as a starting point, the researcher must determine what critical information is lacking that is needed before a simulation can be run.

Comparison of Model to Measurement

The experimental design should describe how the model will be compared to measurements. A good approach is to control sources of error external to the software (inputs) by driving the simulation with measured data (e.g. solar radiation, outdoor temperature) and then to compare the simulated response of the system with the measured response. This may require modifications to the source code and the creation of weather files that contain measured data. The simulation program must be capable of modeling the actual details of the system without making gross assumptions. It is difficult to determine exactly what shortcomings the model may have and how to fix them if the details cannot be examined carefully.

Final Simulation Study

Having verified the accuracy of the model, a simulation study can now be conducted near the range of validation conditions. The modeler should be cautious about extrapolating the model to predict performances under conditions that are significantly outside the range of the validation study; further testing may be needed to verify the accuracy of the model under different conditions. A useful simulation study results in general recommendations regarding the feature or measure that has been studied, including under what circumstances the measure may or may not be desirable.