EFFECT OF A GAS FURNACE CHIMNEY ON THE AIR LEAKAGE CHARACTERISTIC OF A TWO-STOREY DETACIlED HOUSE

C.P. Shaw and W.C. Brown

ABSTRACT

An experimental assessment was made of the effect of a chimney on t he air leakage characteristic of an unoccupfed, two-s t o r e y detached house, as heated by either gas or electric furnace. Measurements were taken of air tightness values and air infiltratkon rates with the chimney capped and uncapped. Infiltration rates were also measured during operation of the elec t r ic furnace.

The over-all air i n f i l t r a t i o n ra te was greater with the gas furnace than with the electric furnace under the same weather conditions, and the d i f f e r e n c e was smaller than the air flow rate through the chimney. When the gas furnace was In operat ion, stack a c t i o n was the dominant: dr iv ing potenria l for air infiltration, with wind speeds up to 7 mJs (16 mph) and with fnside-to-outside temperature dffferences greater than 20 K (36 R).

INTRODUCTION

Because of the r i s i n g cost o f heating fuel and t h e attractive subsfdy from the federa l government for of f -o i l conversion, more and more Canadian homeowners are s w i t c h b g to natural gas for heating. As a conventional gas-fired furnace has a chimney for venting combustion products and there is usua l ly a draft dfverter located near the exft of the combustion chamber, a vertical leakage path extending from basement to roof exfsts in most gas-heated houses. Consequently, i n s i d e air moves upward centintlausly through the chimney during t he heating season. As a result , air infiltration is controlled not only by stack action and wind but also by air movement in the gas furnace chimey.

Although air leakage studies on gas-heated houses have been carrted O u t , ' * the effect of ch imeys on air i n f i l t r a t i o n is still not f u l l y understood. A project was therefore undertaken during the heating season of 1980-81 t o study the a i r leakage characteristic of a two-storey house as heated by either a gas-fired furnace or an electric furnace. The main objec t ives sere to d e t e d n e :

I . the influence of a gas furnace chimney on air tightness and alr infiltration,

2. t h e inf luence of various operating modes of a gas furnace on air infiltration, and

3. the combtned influence of weather and a gas chimney on a i r infiltration.

TEST HOUSE

The house under consideration (Figure 1) is one of four houses involved in the HUDAC/NRC Mark X I energy research project.3 It f s a two- storey dwelling whth a basement, located in a developed residential area in Gloucester, Ontario* It already had an electric forced-air furnace, and for t h i s study a gas furnace w a s installed in parallel with it- Two d i v e s t i n g dampers, one kn the supply and the other in the return af r d u c t , w e r e added to m i n i m i z e cross air flow between the two furnaces- .4 Class C i n s u l a t e d chimney, 12.7 cm ( 5 in.) diameter, w a s installed for the gas furnace . The flue pipe connecting the furnace to the chimney was 10 cm ( 4 in, ) diameter. Whenever khe elec t r ic furnace was in operation, the f l u e p i p e was disconnected from the chimney and a metal cap sealed i n t o t h e chimney openfng. The house was unoccupied when t h e study was conducted.

TEST METHODS

Air Tightness Values

Air t igh tness values were measured with and without the chimney capped, using t h e fan pressurization method. The apparatus4 c o n s i s t e d of

a centrifugal f a n with a rnaximm capacity of 380 LI'S (800 cfm) and a laminar-flow air f low meter (MERIAN LFE Element, accurate to 3% of the measured flow ra te) . A pair of total-pressure-averaging tubes and a thermocouple were installed in t h e f l u e p i p e t o measure t he a ir f l ow rare through the chimney and t h e flue gas temperature, respectively. To f a c i l i t a t e camparison with house infiltration rates, the measured chimney 51m rates were corrected to room temperature.

A i r Infiltration Rates under Controlled Gas Furnace Operation

A i r infiltration sates for the house and air flow rates through the chimney were measured with the gas furnace operating under the fo l l owing conditions: 1) burner on continuously; 2) burner o f f ; and 3 ) burner e y e l i n g under t h e m s t a t i c control.

An automated SF6 constant-concentration tracer gas apparatus5 was used for infiltration rate measurements. The SF6 detector far the apparatus was a pulsed-mode electron-capture detector. To cross-check the results of the SF6 apparatus, infiltration rates were also measured simulraneously by the decay nethod,5 using COp and CH as the tracer gases. Concentra t ions of COZ and CHc were measured with i n f r a r ed and flame-ionization gas analyzers, respectively.

All three detectors w e r e located in the Living room on the ground floor of the house (Figure 1). The tracer gases were injected into the supply duct and samples collected from the return duct through i n d i v i d u a l injection and sampling tubes. Only one furnace f a n was used t o mix the t racer gases with the i n s i d e air.

Air Infiltration Rates under Normal Furnace Operating Modes

Air infiltration rates were measured continuously for three days w i t h the e l ec t r i c furnace I n operation and the chimney capped. For comparison, similar measurements were also taken with the gas furnace in operation. In both instances t h e furnace was under thermostatic cantrol. The infiltration rates were measured with the automated constant- concentrat ion tracer gas apparatus5 only.

Neutral Pressure Level

Neutral pressure levels were measured on a calm day with the burner both on and o f f . When the burner w a s o f f , readings were taken with the chimney capped as w e l l as uncapped. The neutral pressure level w a s determined by drawing a straight l i n e through two simultaneously measured pressure dtfferentials, one across the bottom of an entrance door and the other across the t a p of a second-floor window d i r e c t l y above the door.

RESULTS AN5 DISCUSSION

Figure 2 shows the air leakage rates of the house with the chimney b o t h capped and uncapped. These results were f i t t e d to Equation (1) to obtain the f l o w coefficients and the exponents for both cases:

where

Q = over-all air leakage rate, L / s C c f m ) C = flow coefficient, ~ / s * r n ~ * ~ a ~ (cfm/ft2 (In. of water)" ) A = area of bui lding envelope, area of exter tor w a l l above grade and ceiling area of the top floor, 228 m2 (2454 ft2) bP = pressure difference across exterior w a l l , Pa (in. of water) n = f low exponent.

The f l a w exponent is 0.71 f o r both cases and the corresponding f l o w coef f kc len ts are 0.107 ~ / s - r n ~ . ~ a ~ -71 (1.06 cfm/f t2- inl of wat@rob71) with the chimney uncapped and 0.098 10.97) virh it capped.

The difference between the two air leakage rates is, in theory, the a i r f low rate through the chimney. As shmn in Figure 2, the chimney f l o w rate obtained by t h i s method is smaller than that measured d i r e c t l y . The disagreement is a t t r ibu tab le t o the error of the i n d i r e c t method, which uses the difference between two large air leakage rares to obtain t h e much smaller value for the chimney. Ffgure 2 indicates that for this house the air leakage rate through the chimney constitutes abeur 9% of t h e o v e r a l l air leakage sate.

The air infiltration sates under varfous modes of furnace operation were measured simultaneously, using three d i f f e r e n t tracer gases, Figure 3 compares C02 and CH4 decay results with the SF6 constant-concentration a i r change sates, As s h m , t h e three results agreed closely with one another fox air infiltration rates greater than 0.2 air changes per hour, Belm t h i s rate the SF6 results appeared to be smaller than the corresponding decay results. The difference, however, was just slightly above the experimental error, which was estimated to be about '10% of the measured value.6

FFgure 4 shows the mean values f o r the three simultaneously ob ta ined tracer gas results for the gas burner operating continuously and not operating. The results i nd i ca t e that the infiltration rates w i t h the burner an were, on average, 10% greater than those w i t h it o f f . The difference decreased as t h e inside-outside temperature difference increased.

These infiltsatfon data are compared in Figure 5 with thase measured separately under normal opera t ing conditions o f either t h e gas or the electrfc furnace. The results indicate that f o r s imilar weather conditione the Infiltration rates w i t h the burner cycling were almost identical t o thase with it shut down. When the electric furnace was in

operation (chimney capped), however, there was a noticeable reduction in i n f i l t r a t i o n .

Air flow sates through the uncapped chimney were measured during these tests. As shmn in Figure 6 , a ir flow through the chimney during operation of t he gas burner varied from 22 to 26 E/s (47 to 55 cfm) as inside-outside ternperatuse difference increased from 22 to 42 K (39 .6 to 75 .boR) . Figure 6 also indicates that when the burner was off, the chimey f l a w varied from 18 to 23 L/s (38 to 49 cfrnl. The average f l o w rate w i t h the burner cycling was about 20 L / s ( 4 2 cfm). About 60% of the air exff ltratian occurred through the chimney (Figure 5) ; the remaf nf ng 40% w a s mainly through the c e i l i n g .

The a ir i n f i l t r a t f o n rates of the house with the chimney capped ( F i g u r e 5) were, on average, 11 L/S (23 c f m ) smaller than those with it uncapped. The difference, whfch was much smaller than the average 20 L / s (42 cfm) of air f l o w through the chimney, was due to increased exfiltration resulting from the lavered level of the neutral pressure p l a n e .

Figure 7 shows that wEth an ins ide-outs ide temperature difference of 28 K (50 R ) the neutral pressure l e v e l changed from 3.6 to 4.8 m ( 1 1 . 8 to 15.7 ft), an increase of L.33 times, This caused a flow reversal from e x f ilrsation to id£ f l trat lon a t the exterior wall between the two neutral pressure levels: it increased the infiltration through the lower and middle w a l l s and reduced the e x f i l t r a t i o n through the upper portion of the house. The slope of the pressure d i f f erent ia l profile for t he electric furnace w a s assumed to be Ehe same as that f a r the gas furnace because the s l o p e i s not dependent on the presence or absence of a c h i m e y . Because of lower wind speed, t h e scatter in the data f o r t he gas furnace w a s smaller than that for the electric furnace.

Figure 8 shows that when t h e electric furnace operated normally, t he measured a i r in£ iltration rates for wind s p e e d s lower than 3 . 5 m/s ( 8 mph) agreed closely with the rates predicted by Equation 12) (this equation was derived previously for two houses wirhout chfmneys, inc luded in the Hark XI

I = 0.32 (A/v) C, (~t)" f o r w c 3.5 m / s ( 2 )

where

I = infiltration rate, air changes p e r hour (ac/h) A = area of building envelope, K? ( f t 2 ) v = volume of house, m3 ( f t 3 )

(386 m3 ( 1 3 634 f t3) for t h i s house) Ce = flow coefficlest with chimney capped, ~ / s * m ~ * ~ a "

(cfrn/ft2 (in. of water)n 1 A t = inside-to-outside temperature difference, K (R) n = flow exponent (0.71 for t h i s house).

For comparison, Figure 8 also shows the i n f i l t r a t i o n rates f o r the gas furnace under lm wind conditions. Again, the data were f i t t e d to an expression similar to Equation (2):

I = 0.43(A/v) Cg (~t)" f o r w < 3.5 m l s ( 3 )

where C is the f lm coefficient w i t h t he chimney uncapped and n Fs 0.71. figure 8 indicates that the air infiltration €or this house with the gas-fired furnace operating is abou t 1.5 times greater than that with an electric furnace. The r a t i o of the constants in Equations (2) and ( 3 ) is approximately equal to the ratio of the t w o neutral pressure levels . As earlier s tudies a l so i n d i c a t e that a ir tnfiltration due to s tack action is d i r e c t l y proportional to the n e u t r a l pressure l eve l ,# * 9 a general expression for Equations ( 2 ) and (3 ) would be:

I = 0.32(~/v) r C ChtIn f o r w c 3.5 m / s ( 4 r = H / H

.& e

where X and He are the neutral pressure levels with and without a chimney, r e s p e c t i v e l y . The value of r is 1 for houses without chimneys, b u t can v a r y from 1 t o 2 f o r houses w i t h chimneys. For t h i s house, r is approximately 1.33. C and n In Equation ( 4 ) are measured by means of t h e f a n pressurization method w i t h chimneys, if exfsting, uncapped.

F i g u r e 9 shows the temperature-induced a i r leakage and pressure differential patterns f o r this house with chimney capped and uncapped. The air leakage rates were obtained from Figure 8 f o r an inside-outside temperature difference of 28 K (50 R). The chimney f lm rate was assumed to be 60% of the over-all exfiltration rate (Figure 6 ) . As noted previous ly , the i n f t l t r a t l o n rate with chimney uncapped w a s 50% greater than t ha t : with chimney capped (32 L/S versus 21 L/s). Although over-all cx f l l t ration a l so was greater when the chimney was uncapped, t h e amount of l n s fde air 1eavFng the house through the walls and c e i l i n g was smaller. These resu l t s suggest two reasons why houses without chimneys have mare problems with condensation than those with chimneys:

1) f o r the same tnternal moisture generation rate, houses w i t h o u t chimneys have higher relat ive humidity than t h o s e w i t h chimneys because there is less infiltration for d i l u t i o n ;

2) because there is no chimney venting of t h e i n s i d e a i r chimneyless houses have more humtd air f lowing through the c o l d parts of the upper walls and cef l ing; they are therefore more susceptible to condensation problems.

Unlike e l e c t r i c a l l y heated houses ,7 air in£ iltratdon rates in t h e gas heated dwel l ing w e r e influenced by stack action f o r wind speeds as high as 7 m/s (16 rnph). This suggests t h a t f o r tight houses the strength sf stack act ton is reinforced by the presence of a chimney. Unless, therefore, nFnd s t r e n g t h also increases proportionately, stack a c t i o n remains the dominant d r i v i n g potential for i n f i l t r a t i o n . Measured a i r

infiltration rates under this condition, as shown i n Figure I O , can a l s o be approxfmately expressed by Equation ( 4 ) -

The effect of a gas furnace with a Class C circular chimney 12.7 cut (5 in.) in diameter on the a i r leakage of a two-storey house is as follows:

I . The air leakage value of the house with chimney uncapped is about 9% greater than that with chimney capped.

2. The a i r Fnfiltration rate with the burner on conrinuously is 10% greater, on average, than it is with t h e burner cycling or off .

3. About 60% of i n s i d e a i r leaves the house through the chimney and the remaining 40% exf i l rsa tes through the upper portion of the house envelope.

4 . Switchtng ftout an electric furnace to a gas furnace results in a 50% increase in air infiltration with wind s p e e d s lower than 3.5 mJs.

5. Stack action reinforced by the exhaust air through the chimney remains the dominant driving potential f o r %nfiltratFon with wind speeds up to 7 m / s and inside-to-outside temperature differences greater than 20 K.

REFERENCES

I - Malik, N. F i e l d s t u d i e s of dependence of atr inffltration on o u t s i d e temperature and wind. Energy and Rui ldfngs . - 1 (3) 1977/1978.

2. E l k i n s , R.H. and Wensman, C.E. Natural ventilation a€ modern tightly constructed homes. Presented at t h e American Gas Association Institute of Gas Technology Conference on Natural Gas Research and Technology, Chicago, Ill., 1971.

3 . Quirouette, R.L. The Mark X I research project - Design and construct ion. National Research Councll of Canada, Divis ion of Bui ld lng Research, Building Research Note No, 131, 1978.

4. Shaw, G.Y. and Tamra, G.T. Mark XI energy research project - Air t i g h t n e s s and air infiltration measurements. National Research Council of Canada, Divf s ion of Bu i ld ing Research, Bui ld ing Research Note No- 162, 1980.

5. Rumar, R. , Ireson, A.D. a d Orr, K,W. An automated a i r infiltration measuring system using SFs tracer gas in constant c o n c e n t r a t i o n and decay methods. A S W Transactions, E p a r t 2, 1979, pp. 385-395.

6 . Bassett, H . R . , Shaw, C . Y . and Evans, R.G, An appraisal sf the sulphur hexafluoride decay technique for measuring air i n f i l t r a t i o n rates in bui ldings , ASHRAE Transactions, - 87 part 2, 1981-

7. Shaw, C.Y. A correlation between a i r infiltration and air tightness f ar houses In a developed residential area. ASHRAE Transactions, - 87, p a r t 2, 1981.

8. Shaw, C . Y . and Tamura, G.T. The calculation o f a i r infiltration r a t e s caused by wind and stack a c t i o n far t a l l bui ld ings . ASHRAE Transactions, - 83 part 2, 1977, pp. 145-158.

9. Tamura, G.T. The calculation of house inffltration rates. ASRRAE Transactions, - 85 part 1, 1479, pp, 58-71.

ACKNOWLEDGEMENT

The authors wish to thank B.W. O r r for his guidance in operating the automated constant-concentration tracer gas apparatus. They a l s o wish to acknowledge t he c o n t r f b u t i o n of Dr. D.G. Stephenson in the p r e p a r a t i o n of this paper and t h e assistance of R.G. Evans, J . Payer and D.L. Losan in cocducting the t e s t s .

L I

= = 1 I

I I r* .=- ----:JF---:-------: - _ _ _ _ _ _ - _ _ - - - _ _

FRONT FLNATION

G A 5 FW EI4 ACE

SECTION

C 0 2 DETECTOF

CH4 DETECTOR

TAP

A KNOWN AIR LEAKAGE KATE

{NQl R E ~ l J l I I E U FOK INTILTKATION M E A S I I R F M E N ~ S ~

BASEMENT PLAN GROUND FLOOR PLAN SECOND FLOOR PLAN

FIGURE 1

TEST HOUSE

OVER-ALL

CHIMNEY

UNCAPPED, LL0' --CIMMY

CAPPED

Qe

A:' O/ r

FIGURE 2

OVER-ALL A I R LEAKAGE RATES WITH AND IdTTHOUf CHIMNEY CAPPED AND A I R FLOW RATE THROUGH CHIMNEY

PRESSURE DIFFERENCE ACROSS EXTER lOR WALL. Pa

FIGURE 3

COMPARISON OF A I R INFILTRATION RATE BETWEEN SF CONSTANT CON- CENTRATION AND O2 OR CH4

I DECAY METHODS

f

SFb CONSTANT COYCENTRATION RESULTS. a d h

INS i DE-OUTS I DE TEMPERATURE DIFFERENCE, "K

50

45

40

35

30

25

FIGURE 4

1 -

* - *

* * 3 7

. OQ - a -

W I N D S P E E D < t m / %

a G A S F U R N A C E ON I

AIR INFILTRATION RATES UNDER VARIOUS MODES OF FURNACE OPERATION

M I 1 I 45 CHIMNEY < , > N O -

a / CHIMNEY

20

15

10

- w - ,/ b GAS FURNACE I ( 6 m/s

ELECTRIC FURNACE W <3.5 m/r

I ;c'

- - 0 G A S F U R N A C E O r F .

C H I M N E Y U IJCAPPED

- - 1 C F M - 0 . 4 7 2 LJ;

I " R - 0 . 5 4 ' K

I I I I I I I l l l

l5 t n GAS FURNACE ON CQP4TIMUOWl;LY

0 GAS FURNACE OFF, CHIMNEY UNCAPPED

GAS FURNACE OPERATING NORMALLY

GAS t WkNACE OFF. ELSCIRIC t UKNACt

UN, CEIIMNEY CAWED 1

15 2Q 30 40 5 0 60 70 80

INS lDt-QIJIS IDE TEMPERATURE DIFFEREMCE. "K

FIGURE 5

INFILTRATION RATES UNDER COWTROLLED AND NORMAL FURNACE OPERATION

A I R FLOW RATE THROUGH CHIMNEY UNDER NATURAL CONDITIONS AND RATIO BETWEEN CHIMNEY EXHAUST RATE AND INFILTRATION RATE OF HOUSE

vl 30 -

2

>- LLi

E 20 - I U

L

2

g 10 1 C

5 S u 0

I 1 I I *

l .. . P o - & -

00

w GAS FURNACE ON CONTlNUOW5LY

- - 0 GAS FURNACE OFF, CHIMNEY UNCAPPED

a GAS FURNACE OPERATING NORMALLY

1 I I 0 10 2 0 30 40 5 0

IMSI DE-OUTS I DE TEMPERATURE DIFFEREMCE, O K

AT- 28 K (50 R)

WIND CAM

TOP OF WINDOW ------

-* GAS FURMACE BURNER 06.1

CHlMNEY UNCAPFTD

fPDSS10LE RANGE)

-a- ELECTRIC TURNACE wlTn GAS

FURNACE CHIMNEY CAPPED

IPGSSIBLE RANGE!

FIN. 111 FLOOR

- 6 - 4 - 2 0 2 4

PRESSURE DIFFERENCE ACROSS MTER IOR WALL, Pa

FIGURE 7

NEUTRAL PRESSURE LEVELS UNDER OPERATION OF GAS OR ELECTRIC FURNACE

0.8

0.7 GAS FURNACE

0.6 * W < I ~ J ~ GM FURNACE 0 I G W < Z r n / r

0,5 x 2 4WC3.5 m/5

ELECTRIC FURNACE.

N O CHIMNEY D. 4

a M < 3 . 5 m / r

0.3

0.2

ELECT RlC FURNACE

0 .1 10 20 3 0 40 50 611

I N S I D E - O U T S I DE T E M P E R A T U R E D l F F E R E N C E . " K

FIGURE 8

AIR INFILTRATION RATES FOR ELECTRIC AND GAS FURNACES UNDER LOW WIND

* 2 Po NO CHIMNEY

--- NEUTRAL

PRESSURL

INFILTRATION

1st FLOOR

' - - - _ - - - - - - - - f

/// / BASEMENT

- d - 2 O 2 4 6 8

AP, PRE55URE DIFFERENCE

ACROSS EXTERIOR WALt WE

TO STACK ACT1 ON, Po

CHIMNEY

NEUTRAL

PRES5URE

LEVEL

IrlFLOOR -- ---A-

- 4 - 2 0 2 n a * AP. PRESSURE DIFFERENCE

ACROSS EXTCPIOR WALL DUE

TEMPERATURE-IN DUCED TO STACK ACTION. Pa

AIR FL#l PATTERN

FIGURE 9

TEMPERATURE- INDUCED PRESSURE AND ATR FLOW PATTERNS UNDER OPERATION OF ELECTRIC OR GAS FURNACE FOR A t = 28 K

X S < W < b m / r

A b < W < 7 r n / r

10 2 0 30

I N S IDE-OUTSIDE TEMPERATURE D I F F E R E N C E . O K

FIGURE 10

INFILTRATION RATES FOR GAS FURNACE UNDER MODERATE WIND