Simple error reduction in tracer-gas field-measurements of air handling units

Tracer gas measurements are an unparalleled means of measuring air recirculation, leakage, and air flow rates in air handling systems [1-5]. However, such measurements are subject to significant measurement uncertainty in field conditions. A common problem is imperfect mixing of tracer gas.
One method of error minimization is to use a constrained regression model based upon mass conservation equations. This technique is widely used in multizone tracer gas studies. The method reported in this paper uses robust regression (least trimmed weighted squares) to solve the mass flow rates in air handling units (AHU) for balanced ventilation, including any recirculation/leakage in the AHU, external short-circuiting, and ex-/infiltration caused by imbalance. The method is simple enough to implement in spreadsheet software, and works even when measurement data from one or more of the sampling points, or one of the two dosing points, has to be rejected as erroneous.