Delivering sustainability promise to HVAC air filtration–part I: classification of energy efficiency for air filters.


In 2006, the ASHRAE Board of Directors approved a strategic plan to

lead the advancement of sustainable building design and operation. As

William Harrison, 2008-2009 ASHRAE President, said in his speech

‘Maintain to Sustain–Delivering ASHRAE’s Sustainability

Promise’ at ASHRAE annual meeting in Salt Lake City (Harrison,

2008), sustainability for ASHRAE means energy efficiency and healthy,

productive indoor environments. How can we delivering such

sustainability promise to air filtration in the HVAC system? We have the

standards available in the industry to test general ventilation air

filters for removal efficiency by particle size.

ASHRAE 52.2 and EN779 are the most popular ones in the world with

the former widely used in North America and the latter in Europe. But

none of them has ever addressed the issues on energy efficiency. As

energy becomes a global problem today, there is increasing demand to

establish a method generally accepted to the industry to determine how

to classify a filter’s energy efficiency. This paper is a

preliminary research to introduce two methods which can be used in the

industry for the energy classification. Experiments were made with 6 air

filters (V-pack, bag, box, panel) by ASHRAE 52.2 full test and 7

representative filter media (flat sheets) by TSI 8130 Automated Filter

Tester to demonstrate how to use each method to classify the energy



Method 1: kep Number

Key energy performance number (kep number) is currently being

discussed in Europe for the energy efficiency classification (Mayer, et

al., 2008). It was defined as:

kep = [[-log(1 – [bar.FE])]/[[bar.[DELTA]P] – C]] * 100 Pa (1)

Where [bar.FE] is average filtration efficiency over the dust

loading in EN779 test for a particle size of 0.4 [micro]m @ 3400

[m.sup.3]/hr (2000CFM); [bar.[DELTA].P] is the average pressure drop in

Pa, i.e. Pe, i.e.

[bar.[DELTA]P] = [1/M][M.[integral].0][DELTA]P(x) * dx (2)

M is the final amount of dust loading (g), x is a variable of

loaded dust (g). C is an empirical constant.

The higher the kep value, the higher the filters energy and

filtration efficiency. Thus it classifies all the air filters into 5

levels, i.e. Class 1 to Class 5 as shown in Table 1. Class 1 indicates

the most efficient and Class 5 the least efficient.

Table 1. Energy and Filtration Efficiency Classification of Air Filters

kep Number @ 3400 m3/hr (or 2000CFM) Energy Efficiency Class

kep [greater than or equal to] 1 1

1 > kep [greater than or equal to] 0.8 2

0.8 > kep [greater than or equal to] 0.7 3

0.7 > kep [greater than or equal to] 0.6 4

kep < 0.6 5

A similar method can be used for energy efficiency classification

of the air filters using ASHRAE Standard 52.2 test (Method of Testing

General Ventilation Air-Cleaning Devices for Removal Efficiency by

Particle Size), where [bar.FE] is the average efficiency over the dust

loading at E1 (0.3[micro]m-1.0[micro]m) or E2 (1.0[micro]m-3.0[micro]m)

or E3 (3.0[micro]m-10.0[micro]m) depending on the range of MERV (minimum

efficiency reporting value). FE

Method 2: Wattage

A new method proposed in this research is to directly use the power

of energy (Wattage) which is needed to overcome the flow resistance of

the filters to express the energy efficiency of the filters just like we

rate the power of electric light bulbs. This method treats energy

efficiency independent of filtration efficiency or MERV. For the air

filtration system, the energy consumption (E) can be expressed as:

E = Q * [bar.[DELTA]P] * t/[eta]/1000 (3)

Where E is in kWh; Q is airflow rate in [m.sup.3]/s; [bar.[DELTA]P]

is average pressure drop in Pa; t is operation time (s); [eta] is system

energy efficiency, which is a product of motor efficiency, fan

efficiency, and transmission efficiency. Currently, the state of art

HVAC filtration system can reach the energy efficiencies up to 85%

([eta] = 0.85). The typical value of [eta] is between 0.50 and 0.70

(Mayer et al., 2008). For classification of energy efficiency of a

single filter, the power required to run a filter can be expressed as


W = Q * [bar.[DELTA]P]/[eta] = v * A * [bar.[DELTA]P]/[eta] (4)

Where W is the power in Watts; v is face velocity (m/s); and A is

the face area ([m.sup.2]). Set v = 2.5 m/s or 492 fpm as it is a typical

value used for air handling units and the standard test, we have:

W = 2.5A * [bar.[DELTA]P]/[eta] (5)

Set A = 24″ x 24″ = 0.610 x 0.610 m = 0.3721 [m.sup.2]

for a standard 24″ x 24″ filter and system energy efficiency

[eta]=0.7, then,

W = 1.329[bar.[DELTA]P] (6)

Therefore, we see the key problem to classify the filter energy

efficiency is to find out how to get the average pressure drop during

the use or dust loading process, no matter using kep number method via

Equation (1) or using wattage method via Equation (6)


In reality, pressure drop curves vary with different filters. For a

given filter, it is a dynamic variable which is a function of airflow,

dust type and loading amount, and air conditions as shown in Equation


[bar.[DELTA]P] = [[M.[integral].0][DELTA]P(Q, K, T, H, x(t))/M] (7)

Where Q is the airflow rate; K is a parameter associated with the

type of dust and loading method; T is air temperature; H is air

humidity; x(t) is a variable for the amount of loaded dust; and t is

operation time. To simplify such complexity, the average pressure drop

is normally calculated only based on the initial and final pressure

drop. There are three models used in the industry to calculate the

average pressure drop. They are arithmetic, geometric, and integral.

For the arithmetic model,

[bar.[DELTA]P] = [1/2]([[DELTA]P.sub.initial] +

[[DELTA]]) (8)

Where [[DELTA]P.sub.initial] is initial pressure drop in Pa;

[[DELTA]] is final pressure drop in Pa.

For the geometric model,

[bar.[DELTA]P] = [square root of [[[DELTA]P.sub.initial] *

[[DELTA]]]] (9)

For the integral model,

[bar.[DELTA]P] = [[DELTA]P.sub.initial] +

[1/3]([[DELTA]] – [[DELTA]P.sub.initial]) (10)

All these three models approximate a weighted average of the

pressure drop curve using the initial and final pressure drop; however,

such average pressure drop is subjective to a certain degree as it does

not consider the severity of the curvature which affects the actual

energy performance. Figure 1 shows dust loading curves for three of six

filters in this research tested by standard ASHRAE 52.2 full test. The

shape of the curve was found to nearly perfectly follow the general

exponential equation as expressed below.


[DELTA]P = a * [e.sup.bx] (11)

Where a and b are constants associated with individual filters and

x is the amount of loaded dust (g). Using the initial pressure drop with

a second parameter to describe the curvature one can match the pressure

drop curve during dust loading as seen in Figure 1. The calculated data

from the exponential model showed excellent consistence to actual

pressure drop measured during the ASHRAE 52.2 test. Table 2 lists the

constant values of the exponential model for all six tested filters in

this study.

Table 2. a, b Values of Exponential Models

Filters a b R-squared Value

V-pack Filter 1 79.690 0.0820 0.996

V-pack Filter 2 89.490 0.0056 0.998

Box Filter 55.010 0.0057 0.998

Bag Filter1 44.930 0.0018 0.996

Bag Filter2 36.165 0.0027 0.999

Panel Filter 36.238 0.0027 0.999

Taking Equation (11) into Equation (2), we have:

[bar.[DELTA]P] = [1/M][M.[integral].0]a * [e.sup.bx]dx =

[a/[Mb]]([e.sup.bM] – 1) (12)

Table 3 lists the average pressure drop obtained from different

[bar.[DELTA]P] models. The power rate was calculated based on the

exponential model as it well reflected the experimental data obtained in

the standard ASHRAE 52.2 test with R-squared value in the range of

0.996-0.999. The error introduced by using one of the other 3 models is

illustrated in the Max Error column of Table 3. Compared to the

arithmetic and geometric models which caused the error up to 49%, the

integral model appeared a good model to calculate the average pressure

drop for the filters investigated in this research.

Table 3. [bar.[DELTA]P] (Pa) Calculated by Four Different Models and

Energy Efficiency in Watts

Tested Arithmetic Geometric Integral Exponential Power Max

Filters (watt) Error


V-pack 216.2 146.6 163.2 168.8 220 28

Filter 1

V-pack 226.3 170.5 176.7 185.0 250 22

Filter 2

Box 233.6 185.9 186.5 196.0 260 19


Bag 207.5 115.2 149.9 150.3 200 38


Bag 210.0 129.9 147.7 140.6 190 49


Panel 205.0 114.6 148.3 140.9 190 46


With the power wattage in Table 3, we can report the energy

efficiency (round to 10). Taking V-pack Filter 1 for example, the energy

efficiency is 220 watts, which indicates that the filter requires 220

watts of energy to run at 3400 [m.sup.3]/ hr (2000CFM) airflow, which is

about equivalent to lighting with 4 traditional incandescent light



Electret filters are currently the most energy efficient air

filters on the market (Choi, 2008) as their built-in additional

electrostatic charge greatly improve the filtration efficiency,

especially to capture small particles as shown in Figure 2.


Listed in Table 4 are seven air filter media (flat sheets) from the

market, where four were pure mechanical and three were electret charged.

The filtration efficiency and pressure drop were measured on a TSI 8130

automated filter tester at airflow 48 l/min, using NaCl aerosols. The

kep number was calculated according to Equation (1), here [bar.FE] was

the average of three efficiency readings directly from the TSI 8130. The

power wattage was calculated according to Equation (6).

Table 4. Different Filter Media Energy Efficiencies

Sample Filtration Pressure kep Energy Power

Efficiency drop number Efficiency (watt)

(%) (Pa) Classification

Glass fiber1 87.2 850.1 0.11 5 1130

Glass fiber2 98.3 970.2 0.18 5 1289

PP 87.3 910.0 0.10 5 1209

Polyester 48.5 138.5 0.21 5 184

Charged PP1 97.3 42.5 3.68 1 56

Charged PP2 99.3 55.6 3.84 1 74

Charged Polyester 53.8 13.4 2.01 1 18

* Calculation was based on Equation (1), here C=0

From the power data in Table 4, electret media showed significant

advantage in energy efficiency over glass fiber media and non-electet

synthetic media. The challenge of electret media is potential

electrostatic decay during real use. Thanks to new technologies

developed in recent years that greatly improved the stability of

electrical static charge (Janssen et al., 2006; Tsai, 2003; Motyle et

al., 2006), many eletret filters showed no notable performance decay of

particulate removal efficiency in the field while providing lower

pressure drop with higher energy efficiency over the course of their

service life [Sun, 2008]. Electret filters have gained significant

market share and are expected to grow in air filtration area as energy

efficiency becomes an increasing public concern (Wang, et al., 2008).


Energy cost and use have become a global concern. There is a strong

demand today from end users and filter manufactures to add energy

efficiency to filter classification which is solely based on particulate

removal efficiency in current standards. Two methods were discussed in

this paper: kep number and wattage. The wattage method proposed in this

paper directly uses the power of energy needed to overcome the flow

resistance of the filters to express the energy efficiency of the filter

in a standardized test. Four different models were discussed to

calculate the average pressure drop, including arithmetic, geometric,

integral, and exponential. The exponential pressure drop model proposed

in this research was found excellent to reflect the actual pressure drop

along dust loading in standard ASHRAE 52.2 test, while it caused up to

49% error in power estimates when arithmetic and geometric models were

used. Further filter media investigation using TSI 8130 automated filter

tester demonstrated that electret media are currently the most energy

efficient filter material on the market over non-electret ones.


Choi, K.J. 2008. Energy Efficient Air Filter Media, INTC08,

Houston, TX, Sept. 8-11.

Goel, M. and T. Goel. 1999. Future Trends of Electret Applications

in Energy and Environment, 10th International Symposium on Electrets.

Harrison, W. A. 2008, Maintain to Sustain – Delivering

ASHRAE’s Sustainability Promise, ASHRAE Meeting, Salt Lake City,

June 21-25.

Janssen, L. and J. Bidwell. 2006. Performance of Four Class 95

Electret Filters Against Diesel Particulate Matter, Journal of the

International Society of Respiratory Protection, Vol. 23, Spring/Summer.

Mayer, M., T. Caesar, J. Klaus. 2008. Energy Efficiency

Classification of Air Filters, 10th World Filtration Congress, Leipzig.

Motyl, E., B. Lowkis. 2006. Effect of Air Humidity on Charge Decay

and Lifetime of PP Electret Nonwovens, Fibers & Textiles in Eastern

Europe, Vol. 14, No. 5(59), January/December.

Sun, C. 2008. Mechanical or Electret Filters? INTC08, Houston, TX,

Sept. 8-11.

Tsai, P. 2003. Novel Methods for Making Electret Media &

Remediation of Charge Degredation,” INTC 2003, Renaissance

Harborplace, Baltimore, Maryland, September 16-18.

Wang, A. and K. C. Hofacre. 2008. Assessment of Advanced Building

Air Filtration Systems, EPA/600/R-08/032,, March.

Christine Sun, PhD


Christine Sun is R&D Manager and Dan Woodman is Director of

Engineering, Freudenberg Filtration Technologies, L.P., Hopkinsville,