Several experiments have been conducted to verify the model predictions. Most of these tests have been conducted in the EPA IAQ test house. Tichenor et al. (1988) provide a full description of the test house. The floor plan of the IAQ test house is shown in Figure 3.
Two types of experiments have been conducted in the IAQ test house. One type of experiment used tracer gas with known emission rates. These experiments test the model's ability to track pollutant transport in the building and to respond to changes in factors such as room-to-room air flows (due for example to heating and air conditions (HAC) system operation) and air exchange with the outdoors. The second type of experiment uses real sources and tests the source and sink models as well as the transport portions of the model. The results of the two types of experiments will be briefly discussed.
ASTM has developed guidelines for comparing model predictions with data. (ASTM, 1991). The ASTM criteria include:
(3)
where Cp is the predicted concentration and Cm is the measured concentration. The bars indicate average; e.g., is the average of all the predicted concentrations. The NMSE has a value of 0 when there is perfect agreement for all pairs of measured and predicted concentrations. NMSE is near 0.25 for differences between measured and predicted of about 50%. ASTM recommends that the value of NMSE for an adequate model be less than or equal to 0.25.
(4)
FB ranges from -2 to +2 with a value of 0 indicating perfect agreement. ASTM recommends that the absolute value of FB for an adequate model be less than or equal to 0.25. ASTM recommends that the model be judged based on all the criteria. It is possible to have an adequate model even if all the criteria are not met.
More: