Long known to correlate with human metabolic activity, carbon-dioxide (CO₂) levels are a reliable indicator of indoor pollution originating from building occupants. Demand-controlled ventilation (DCV), the process by which outside air is varied based on the amount of CO₂ in a space, has been around for 15 to 20 years. Yet despite the energy-saving and health benefits of providing just the right amount of outside air needed by building occupants, DCV is underutilized because of concerns about non-human indoor pollutants,¹ as well as issues surrounding control accuracy, sensor calibration, and maintenance. This article explains how a centralized facility-monitoring system capable of sensing multiple points and both human and non-human pollutants — multi-parameter demand-controlled ventilation (Mp DCV) — resolves all of those perceived limitations.

SENSOR PLACEMENT

With a conventional DCV system, discrete CO₂ sensors are installed in the spaces to be controlled. These sensors generally report to a building-management system (BMS) or, in the absence of a BMS, directly control outside-air dampers or variable-air-volume boxes. Because ANSI/ASHRAE Standard 62.1, Ventilation for Acceptable Indoor Air Quality, is concerned with the difference between indoor and outdoor, rather than absolute, CO₂ levels, sensors for outside air also must be provided. (As Figure 1 shows, outside-air CO₂ levels can vary greatly over a 24-hr period.)

Mp DCV replaces as many as 30-plus discrete sensors in individual spaces with a single centrally located sensor. Air samples are transported from an outside-air sampling point and each space being monitored to the centrally located sensor, most commonly through the use of small-bore flexible conduit (microduct), with the sequencing of the samples controlled by electronically operated solenoid valves.

With a typical air sample having a volume of approximately 0.5 cu ft and a velocity of 20 fps, microduct must be highly conductive to prevent the buildup of electrostatic charge on particles in the air. It also must be extremely inert to prevent sorption or desorption of sample constituents.

Because the number of sensors is reduced greatly, compliance with the American Society of Heating, Refrigerating and Air-Conditioning Engineers-recommended calibration frequency of every six months is much less costly with Mp DCV than it is with conventional DCV.

SENSOR ACCURACY

Most commercial- and laboratory-grade CO₂ sensors are based on non-dispersive-infrared (NDIR) technology — that is, an infrared (IR) light source and an IR detector. As with any light source, accuracy varies, and devices fail. With Mp DCV, the number of devices to be replaced is reduced greatly, resulting in significant cost savings and fewer disruptions of operations.

The tendency of NDIR sensors to drift gives rise to concerns about control inaccuracy. Most sensors have accuracies of ±75 ppm. If an outside-air sensor and a single inside space sensor were to drift the same amount in the same direction, the errors would cancel, and the desired CO₂ differential would be maintained. The worst-case scenario for sensor error translating to ventilation error would be for both sensors to drift the maximum amount in different directions. Figure 2 shows CO₂ ranging from 375 to 675 ppm. In the former case, the space is significantly underventilated because CO₂ is thought to be 150 ppm below the desired threshold; in the latter case, the space is overventilated because CO₂ is thought to be 150 ppm above. That is a +29-percent error, which translates to approximately 6 cfm of unnecessary ventilation.

To the owner of a building with 500 full-time-equivalent occupants paying 10 cents per kilowatt-hour for electricity and $5 per cubic foot per minute per year for conditioned outside air, the cost in wasted energy is $15,000 annually. Although this is a simplistic — and somewhat conservative — example, it indicates the magnitude of the savings that can result from improved ventilation. The environmental impact is approximately 92 metric tons of CO₂, which, according to the U.S. Environmental Protection Agency's (EPA's) Emissions and Generation Resource Integrated Database (eGRID) (www.epa.gov/cleanenergy/energy-resources/egrid/index.html), is the equivalent of 18 average automobiles burning nearly 11,185 gal. of gasoline.

In 2006, Lawrence Berkeley National Laboratory undertook a pilot study of commercial buildings in which CO₂ sensors were installed.² Although the relatively small sample size (nine buildings) provided only an initial indication of the in-situ performance of the sensors, the study indicated a need for more accurate CO₂ sensors and validated the need for better maintenance and calibration.

Acceptable Use Policy blog comments powered by Disqus