The following is a summary of an LBNL report on CO2 monitoring for demand controlled ventilation in commercial buildings
On March 17, 2010 the Lawrence Berkeley National Laboratory (LBNL) released a report entitled, CO2 Monitoring for Demand Controlled Ventilation in Commercial Buildings, which concluded “that many CO2 based demand controlled ventilation systems fail to meet the design goals of saving energy while assuring that ventilation rates meet code requirements.” The primary reason for this failure to achieve the design goals of DCV is due to the fact that the accuracy of the single-location CO2 sensors are frequently less than adequate to meet these goals. Specifically, the report mentions that when tested at 760 ppm of CO2, 47% of the sensors had errors greater than ± 75 ppm and 31% of the sensors had errors greater than ±100 ppm. In addition to this equipment problem with the sensor accuracy, there was also the administrative problem due to the fact that none of the facility managers in this study had indicated that they had calibrated any of the sensors since the time of sensor installation.
While this document also reports that there is good justification for monitoring CO2 concentrations and using these values to modulate rates of outdoor air delivery, this operational strategy can only be expected to be effective if CO2 sensors have a reasonable accuracy in practice. One option for achieving this accuracy is to use multi-location monitoring systems as opposed to single location sensors. These multi-location monitoring systems have their advantages and disadvantages. The advantages include the use of one sensor to measure CO2 values at multiple locations potentially reducing total sensor costs, the potential to spend more to obtain a higher quality sensor with its cost amortized over multiple locations, the ease of calibrating a single or small number of sensors relative to calibrating many sensors, the potential to include an outdoor air CO2 measurement in each building, and the potential to simultaneously monitor other IAQ-related parameters, such as absolute humidity, along with the CO2. Disadvantages include the need for a multi-location sampling system of tubing, valves, and pumps, with the potential for leakage-related errors with multi-location sampling, and the reduced frequency in which CO2 concentration data are available from each location.
Another advantage of the multi-location sampling approach is that the ventilation rate determination is based on the difference between the indoor and outdoor CO2 concentrations, so that sensor offset errors are cancelled out, thus achieving a much higher degree of accuracy as compared with either using two sensors or merely relying on the absolute value of the single indoor sensor.
More details of this LBNL investigation, including the results that suggest that measurements at return air grilles may be preferred to measurements at wall-mounted locations, can be found at:
http://www.demandcontrolledventilation.lbl.gov/pdf/lbnl-3279e.pdf
David W. Bearg, PE, CIH IEQ is a building scientist with Life Energy Associates