October 2004 
Innovations in Comfort, Efficiency, and Safety Solutions. 

Leonard A.
Damiano, 
SUMMARY
The minimal installation labor requirements and the simple setpoint control concepts involved in using CO2 are so seductive, that normal caveats are ignored leading to misunderstandings of the principles, misapplications and misuse of the methodology. We need to start recognizing the verified limitations of demand controlled ventilation (DCV) and apply it only where appropriate and where no damage or harm could occur.
INTRODUCTION
There are many ways to use CO2 measurement as the primary or secondary input to automatically control outside air intake rates. CO2 measurement is principally used to estimate the number of occupants in a defined space, based on the mathematical relationship between CO2 respiration and ambient outdoor rates, which are then used to calculate the estimated amount of outside (ventilation) air being supplied to the space.
Because of the added instrumentation costs and offsets by the energy savings expected in variable occupancies (compared to continuous supply of minimum rates for max ), CO2 –based DCV should apply only to dense, unpredictably variable and intermittent occupancies.
Five of the more popular nonproprietary ways include:
Control ventilation directly from interior CO2 concentrations alone;
CO2 Mass Balance;
CO2 Steadystate Concentration Balance (ref. ASHRAE 621999 Appdx. D / 2001 Appdx C),
OA Measurement with Reset by DCV (using CO2 or other occupancy input); and,
DCV Control only between Upper/Lower OA design limits, established by direct measurement.
DCV direct control has been attempted using only an indoor concentration measurement, with poortomixed results. Interior concentrations alone have no direct relationship to ventilation rates.
Most methods require CO2 input as the differential (net indoor) concentration calculated from indoor CO2 minus outdoor CO2. This method provides an approximation of human occupancy based on many assumptions, including zero measurement error, constant respiration and production rates. Continuous outdoor CO2 measurement is required as one prerequisite in the conditional use of DCV for compliance with the Ventilation Rate Procedure of ASHRAE 62 and CEC’s Title 24 ventilation code.
CO2 –based DCV is not explicitly allowed by either currently used versions of the International Mechanical Code2000/2003. In fact, several requirements in IMC Section 403 combine to effectively prevent its broad application.
The authority that many have used to justify CO2 measurement for direct control of ventilation rates, originates with the CO2 SteadyState “Concentration Balance” Equation, referenced in Appendix C of ASHRAE Standard 62 – 2001. Even though the appendix is explicitly provided as informational only and not part of the standard, the concept has been pirated to support of CO2 –based DCV. It also appears that every packaged equipment manufacturer in the world that offers a basic controls package has its own unique twist on how “best” to use CO2 inputs. Because this indirect measurement can never truly indicate intake rate deficiencies, equipment suppliers may embrace this method to dodge the bullet that would otherwise verify or refute the intake performance and capabilities of their products.
Please note that none of these methods can be used to estimate timespecific ventilation rates, as referred to in ASTM Standard D624598, Guide for Using Indoor Carbon Dioxide Concentrations to Evaluate Indoor Air Quality and Ventilation. This is a very specific procedure using a single instrument to measure both multiple inside and outside concentrations, with a great amount of research behind it and the one still others have mistakenly used to justify direct ventilation control using CO2.
STEADYSTATE OR CONCENTRATION BALANCE
First, examine the relationship between the mathematical components of the Steadystate Concentration Balance equation.
ANSI / ASHRAE Standard 621999 and 2001 provide us with the SteadyState Concentration Balance Equation (below):
V_{O} = 
N 
(C_{s}  C_{o} ) 
Where,
V_{o}
= Outdoor air flow rate per person
N
= CO2
generation rate per person
C_{s}
= CO2
concentration in the space
C_{o}
= CO2 concentration in outdoor air
The equation can be converted to volumetric units of CFM and concentrations in ppm. The revised equation becomes:
Vo = 10,600 / (Cs  Co)
When applied to 1,000 ppm set point (700 ppm above a fixed outside base of 300 ppm); Vo = 10,593 / (1,000 – 300) = the familiar 15 CFM/person. If we calculate based on 800 ppm set point (differential of 500 ppm), the result is 21 cfm/person. Carrying this to an extreme, we can calculate for 600 ppm and get a required ventilation rate of 35 CFM/person, while remembering that the outside base was held constant at 300 ppm.
If we look at current CO2 levels in cities, we find that an average closer to 400 ppm. Readings above 500 are not unusual. Los Angeles and Mexico City have reported readings of 600 ppm or better. This puts a great burden on those who insist on controlling indoor CO2 to a specific maximum level, especially those that assume “if some is good, more is better”. For example, if we had an outdoor CO2 level of 450 ppm and attempted to control ventilation to an indoor level of 600 ppm, we calculate that 70 CFM/person is required. Not very reasonable, is it?
A minority of authorities feel that increasing the amount of dilution ventilation based on a specific indoor CO2 concentration may be assumed by some to be “linear and absolute”, when in fact it has been shown to be “inverse and relative”. ref. Feber, T.R., ASHRAE Journal. Lowering the internal level of CO2 beyond a certain point does not necessarily provide the positive, energy saving results that are desired.
EFFECT OF CO2 MEASUREMENT ERROR
Based on the ASHRAE Appendix Steadystate Concentration Balance equation, if we conclude that the ventilation rate required in a space is to be 15 CFM/person, we can calculate the CO2 differential (inside to outside) to be 689 ppm, when N=0.31 L/min (equal to the respiration of a seated person). The calculation using 20 CFM/person as our objective provides a CO2 differential of 517 ppm, using the same parameters.
These calculations must assume that measurement accuracy is perfect – no error, which is not realistic. What happens when our measurement of CO2 contains error (% Reading)? What impact does the error in CO2 measurement have on outside air intake rates being controlled?
Holding the outside CO2 concentration constant, a straight calculation of the CO2 differential including error rates from +2% to +10% of reading provides us with intake/person errors of –3.1% to 13.6%. As the positive error of outside CO2 measurement increases, the intake rate/person decreases.
EFFECT OF OUTSIDE CO2 VARIABILITY
We know that outside CO2 is not static and that it varies both geographically and over time, throughout a single day, as well as seasonally. Therefore, making the assumption that a single outside concentration level is valid continuously is not a very professional engineering practice.
To determine the potential impact of external variations in concentration on ventilation rate control, we make the same calculations, but this time vary CO2 from 300 ppm to 500 ppm. This range is also fairly typical. The resulting calculation tells us that the intake error rate per person will range from –12.7% at the lower CO2 rate to +17% at the high end of this example.
In another example, if we had a local outdoor CO2 level of 450 ppm and attempted to control ventilation to an indoor level of 600 ppm, we calculate that 70 cfm/person is required to offset this exaggerated and minimal CO2 differential.
EFFECT OF APPLYING MULTIPLE CO2 SENSORS
The various methods of using CO2 measurement are not the same (DCV, Mass Balance, ASHRAE Concentration Balance and ASTM tracer). All of these methods, when compared to DCV, are timespecific analytical tools that use the insidetooutside ratio of CO2 concentrations, together with the internal generation rate of carbon dioxide per person, to calculate estimates of the amount of ventilation air that is available per person in the space. DCV attempts to modify the intake rate based on changes in occupancy, in this case, as a result of changes in CO2 concentrations. It is important to also recognize that these methods assume the use of a single measurement instrument to help validate their calculations.
DCV, on the other hand, is a method attempting to control ventilation intake rates for a building (or space) based on the internal CO2 concentration, many times by averaging multiple measurement points internally (or the amount in the return duct system) and comparing that to an assumed average value or a single outside measurement.
MASS BALANCE
An assessment of CO2 controlled ventilation was recently expressed in the evaluation of the “mass balance” technique of calculating outside air intake rates, as described in the contents of ASHRAE Research Project RP980 (1999). Error Analysis of Measurement and Control Techniques of Outside Air Intake Rates in VAV Systems, conducted at UCBoulder, Department of Civil, Environmental, and Architectural Engineering. We quote from the report.
The concentration balance airflow measurement technique [Author’s Note: actually “mass balance”, see below]…. is performed using one sensor to measure all three CO2 concentration values [Recirculation air, Supply and Outside air]. Using multiple CO2 sensors to determine the outside airflow rate is not possible due to the relatively large error associated with the absolute accuracy of commonly available sensors.
….. However, this requirement for relatively stable CO2 concentrations limits the applicability of the concentration balance technique. In spaces where large, abrupt changes in occupancy (and, hence, CO2 levels) can occur, this method may prove unreliable. This fact may rule out the use of this control strategy in spaces such as conference rooms and auditoriums or any building where large transient effects are possible. …
…. The predicted errors indicate that the concentration balance airflow measurement technique may be valid except when occupancy is low or when the difference in the recirculated and outside air CO2 concentration levels is small. Additionally, when the outside air represents a small fraction of the total supply air provided, errors in the calculated outside airflow may become too large for reliable and accurate use.
The formula used for their “concentration balance” equation is actually a mass balance formula, and is not the same as that referenced in ASHRAE Standard 62 Appendix C.
V_{OA} = V_{SA} ( 
CO_{2RA} – CO_{2SA} 
) 
CO_{2RA} – CO_{2OA} 
where
CO_{2OA} = Outside air CO_{2} concentration, ppm
CO_{2RA} = Recirculated air CO_{2} concentration, ppm
CO_{2SA} = Supply air CO_{2} concentration, ppm
Logically, the uncertainty of using multiple instruments must consider the combined error rates of all the instruments used. This would apply to their use in the “concentration balance” method as described above or their use in controlling intake air based on internal CO2 measurements using the ASHRAE Concentration Balance Equation. Neither provides the mathematical results expected.
Going back to the Concentration Balance formula again, let’s examine the results of instrument error, based only on one measurement. What happens when we calculate for a 0%  10% opposition in uncertainty between the outside and inside CO2 measurements? If we hold the error rate equal at either the inside or outside concentrations, we find that the resulting intake rate error ranges from –3.4% to –15.1%, as indicated in the table below. As the CO2 sensor error rate grows, the intake error increases negatively, about twice the single instrument rate.
If we only measure indoor concentrations with the outdoor component assumed to be constant, we would be wrong during some part of the day or season. This 40% error band really happens. Unfortunately, it is rare that direct intake rate measurements are available in realtime to compare to these calculated estimates. There seems to be a selfsustaining aversion to verification that bolsters the use of this methodology.
The data calculated in the examples above are the measurement errors for a single instrument. The compounding effect of multiple instrument error rates allows much greater uncertainty in the outside air control application.
EFFECT OF VARIABLE RESPIRATION RATES
The final variable that we will examine is “N”, the respiration rate, CO2 generation and their effects on ventilation. Is it realistic to assume that everyone in a building will be seated, is of the same size, sex, health, and consuming the same diet? Can all the other factors that influence respiration rates be held constant?
When we examine the range of activities and their impact on respiration, we find that “N” can easily vary to 0.50 for walking, 0.60 for Light machine work and 0.90 for the upper threshold of “light activity”. All of these are cited in the Appendix to ASHRAE Standard 622001. These figures assume a target ventilation rate of 15 CFM. When the target is 20 CFM, then the range of “N” varies from 0.40 for office work, to 0.50 for Walking. When these are used to calculate the amount of ventilation per person, we come up with the following range of rates.
Respiration Rate 
Actual Outside Air Rate 
Activity 

N (L/min) 
L/sec 
CFM / person 

0.30 
7.3 
14.5 
Seated 
0.40 
9.7 
19.4 
Office Work 
0.50 
12.1 
24.2 
Walking 
0.60 
14.5 
29.0 
Light Machine Work 
0.90 
21.8 
43.5 
Upper threshold of “light activity” 
The point is that any change from the minimum respiration and intake ventilation target has dramatic impacts on the amount of ventilation required, if 700 ppm above a constant outside CO2 rate is used.
ASHRAE Standard 622004 (including addendum “n”)
The real crunch for the blanket usage of CO2 –based DCV happened when addendum “n” to Standard 62 was approved. The new Ventilation Rate Procedure changes the basis of determination for the minimum rates used in the tables, from occupancy only to a combination of occupancy and floor area. This added component (floor area) has no relationship to CO2 and nearly establishes a requirement for a base ventilation rate, independent of occupancy.
OA = (Occupants x CFM/person) + ( Area ft^{2} x CFM/ ft^{2})
Here is an example of how the CFM/person requirement changes with Addendum n for an office space at various population levels. If the OA required = 5 CFM/person + 0.06 CFM/sq.ft. Assume a fixed area of 1,000 sq.ft. Then vary the occupancy to see what happens.
No longer can a fixed CO2 concentration (ppm) be used for set point control and equated to an assumed ventilation rate, for compliance with ASHRAE 62 or those ventilation codes that reference it.
CONCLUSIONS AND COMMENTS
Some conclusions about CO2 sensor technology and applications were made from a NIST review of current literature on the subject (Persily, A. and Emmerich, S.J., 2001. StateoftheArt Review of CO2 Demand Controlled Ventilation Technology and Application. NISTIR 6729, National Institute of Standards and Technology). Some of the relevant ones include the following:
The greatest savings are likely to occur in buildings with large heating or cooling loads and with dense and unpredictable occupancies.
DCV may not be appropriate in mild climates.
Avoid DCV in spaces with significant sources other than people.
Avoid buildings with CO2 removal mechanisms.
Both NDIR and photometric detection can be affected by light source aging, NDIR by particle buildup and photometric by vibration or atmospheric pressure changes.
Consideration must be given for selecting only for the appropriate range of operation
Drift is still an issue and calibration recommendations must be followed.
The preferred locations for sensors are multiple ones placed in the occupied zones
Do not use sensors that are not intended for control purposes.
Do not use sensors near doors, windows, intakes or exhausts, or in close proximity to occupants.
Single sensors in the return air should not be used for multiple spaces with very different occupancies.
Economizers should be allowed to override DCV.
Higher outdoor levels of CO2 will result in over ventilation when levels are assumed (not measured) and an outdoor sensor may be required by applicable standards or codes.
We would be reasonable to conclude from the limitations and implications discussed here, that:
DCV has a place and is needed for greater sustainability in construction
DCV offers efficiencies in the operation of unpredictably intermittent and variably occupied structures/spaces and its applications should be limited to those spaces
DCV does not provide the control reliability of other methods of control and is not a suitable method of controlling intake rates for the vast majority of structures.
DCV can be described as “outdoor air control reset due to changes in occupancy”
DCV can be implemented with automatic control inputs other than CO2 reflecting changes in occupancy
DCV can be implemented with CO2 if the occasions for over and under ventilation can be avoided.
DCV can allow the intakes to close completely, helping to provide the energy savings which has been promoted, but which also should not be allowed.
DCV does not currently provide a clear means for compliance with most ventilation codes and standards, except in specific situations.
Conflicts between energy code requirements for CO2based DCV and the deficiencies identified here, need to be harmonized.
RECOMMENDATIONS
Dilution ventilation for acceptable indoor air quality requires that the needed amount of air is actually delivered to the breathing zone of the occupants. This requires attention to the design of controls and air distribution systems, as well as to the type of instruments used for control inputs.
The best controls and air distribution system are no more than useless hardware if the inputs used are not valid, or are so unreliable as to make them ineffective.
What can you do to avoid these problems?
Avoid indirect methods of ventilation measurement and control. Use Demand Controlled Ventilation when appropriate – but with occupancy and airflow rate determinants other than CO2 (e.g. direct velocity measurement, time scheduled occupancy, direct count by turnstiles, IRbased sensors, monitored equipment or lighting usage, etc.).
Design capacity should consider maximum occupancy at “worstcase” conditions.
Establish and insure a base (minimum ventilation rate) with direct intake measurement, regardless of occupancy level.
Dynamically monitor occupancy for changes and reset operating set point as needed.
Look for instrumentation and control equipment justifications beyond their firstcost.
Do not accept ineffective measurement technologies or control methods – “garbagein / garbageout”
The best and most secure method of outside air intake control is direct measurement of intake air rates. With a direct, realtime input any number of HVAC strategies can be realized to optimize energy usage, through dynamic control under changing internal and environmental conditions.
Many excuses have been levied for the avoidance of direct outside air measurement, the most pervasive being cost and the difficulties found in applying traditional technologies to field conditions. But, these excuses are countered with several products and methods that provide both costeffective means and reliability in performance, when used for outside air control.
Airflow measurement products are available for commercial HVAC applications with installed costs comparable or favorable to traditional technologies, and when installed costs are compared to several methods of CO2 control.
A number of other advantages are available to the enduser, including permanent factory calibration to NISTtraceable air speed standards, and with one product in particular, minimal duct placement limitations.
For most applications with minimal variability or lower densities without fixed seating, why risk the energy and liability consequences of indirect measurements, when direct control of outside air intake rates can be accomplished (in most situations): more economically, more effectively and more reliably?
[Home Page] [The Automator] [About] [Subscribe ] [Contact Us]