| Boiling Point No. 28- August 1992 |
by M E Crowther, Domestic Manager, Coal Research Establishment, Cheltenham, England, UK
Recent studies have highlighted the health problems associated with smoke from burning biomass upon stoves and open fires in houses where there is no chimney or other way of directly discharging this smoke to the outside (ref. 1). This problem is greatly increased if the solid fuel used is of fossil origin and has a significant tar or sulphur content. It is universally agreed that this smoke needs to be discharged outside the house, ideally at or above roof level. One method by which this can be achieved is to replace the existing fuel burning arrangement with a closed stove and flue pipe. This approach is ideal for communities that can afford sophisticated and accurately manufactured designs in steel, cast iron or ceramics. Unfortunately low cost chimney stoves not made to these standards, whilst lowering smoke levels in the home, can dramatically reduce efficiency and increase cooking times; thus they cannot provide all the benefits that are desirable.
A popular approach in many African and Asian countries is to dispense with a chimney on the grounds that it's costly and gives problems with operation and maintenance. This means that simple stoves, even with well insulated fire boxes and carefully designed pot supports giving excellent heat transfer characteristics, release the emissions directly into the room when used indoors.
Theoretical analysis of a chimney or hood
There is only one mechanism by which a hood or chimney can effectively discharge smoke outside the house without the use of a fan. This is to employ the natural buoyancy of the hot gases produced during combustion. This has the added advantages of discharging the smoke at high level. A collection hood is mounted over the fire and connected to a chimney of sufficient diameter and height to evacuate the products of combustion. The driving force behind this is the lower density of these flue gases producing a pressure drop up the chimney.
If the chimney is too short, too small in diameter or if there is too little heat released from the fire, insufficient volume of gases will pass up the chimney to collect the smoke and stop spillage to the room. For simplicity of analysis, a smoke hood is best envisaged free standing (Fig1) or against a wall (Fig 2). The important parameter is the total free face area A across which smoke from the fire can spill into the room. For Fig l this is "2h(a+b)", and for Fig 2 this is "ha". This spillage will occur by diffusion and (this is probably more important in a real situation) by minor air currents in the room, caused by movements or doors and windows. There needs to be a certain minimum "Face Velocity" across area A to keep the smoke contained in the hood.
Having set this velocity the chimney design parameters (diameter and height) can be established. In the UK a minimum value for this velocity is usually taken (for fumes such as the smoke from domestic fires) as between 0.2 and 0.4 m/s. (ref. 2). Below 0.2m/s smoke spillage is known to increase particularly with large values of A.
It car, immediately be appreciated that large open hoods mounted some distance above a fire require huge volumes of air to be moved for them to be effective. To quantify this, a free standing hood 1m above the hearth of the fire with a perimeter of 3m would require an air clearance rate (to maintain a face velocity of 0.3m/s) of 3240m³/hr, equivalent to a velocity of 12.2rn/s in a 250mm circular chimney (a large volume even by the standards of a modem extractor fan and one that would require a theoretical power input of 67 watts to overcome the "velocity head"). Mounting the hood against a wall and fitting side wings immediately reduces this to a more tractable, but still large, 540m³/hr.
It is now possible to establish a mathematical model of the system. A chimney over an open combustor operates in dynamic equilibrium: the buoyancy effect of the heated gases just equals the velocity head (the energy required to accelerate the gas to the velocity in the chimney) plus frictional losses associated with the flue gases ascending the chimney.
A set of equations may then be established to evaluate flue gas rate, for a certain chimney diameter, height and heat release rate. These may be solved iteratively with a computer. The programme written to date assumes isothermal chimney operation. It would be possible to modify the programme to model chimneys where heat loss could be significant, for example thin wall metal chimneys. To assist with generating the data for this paper, such a programme has been written in BASIC; copies of this (which are Copyright with the Coal Research Establishment) are available from the author.
Although it was categorically stated above that flue gas buoyancy is the only reliable method of smoke evacuation, other mechanical devices are conceivable, ea. wind driven cowls, but by definition they are dependent upon the wind and thus will only be effective on a certain number of days per year and/or at certain times of day. A will be shown later the pressure drop across a chimney is often not greatly in excess of the "velocity head" generated by the wind and hence operation of short chimneys is critically dependent upon local topography and wind conditions. This paper will establish design conditions for chimneys that are as far as possible independent of such external influences. We believe only chimneys/ hoods built to such specification can be built with confidence for universal application.
Accuracy of the approach
Figure 3 gives predicted flue gas flow rates for 200mm and 225mm square and round chimneys plotted against net or sensible heat release from the fire to the flue gases. A number of measured values are also reported. These measurements were carried out in a laboratory test chimney constructed in accordance with BS3376 (ref. 3). It is 225mm square and 4.5m high (from the floor). The chimney is considered to have an effective height of 4.2m. Agreement can be seen to be generally good. Velocities in the flue are measured in the range 2.0 to 2.3m/s (actual), this equates to operation inside the rectangle shown (ref. 4). A particularly accurate measurement has been recorded at 336Nm/hr (ref. 5). The lower data set (ref. 6) is interesting as it shows a very similar slope for the curve "volume flow rate vs heat release" to that predicted.
Predicted flue gas temperatures which are in the range 68ºC (4kW to chimney) to 99ºC (8kW to chimney) are also in agreement with measured values.
Use of model in chimney/hood design
It is now assumed that the model we have produced is a reasonable reflection of a real chimney, it can be used to calculate volumetric flows that are produced for a given heat release rate. For convenience this data is plotted (Fig 4) as volumetric flowrate against chimney diameter at 2kW and 6kW (left hand scale). On the right hand side is the corresponding free face area at a velocity of 0.3m/ s (by coincidence this is numerically equivalent to dividing the left hand scale by ten). Because this paper is concerned with reducing smoke to room it is appropriate to use the right hand scale
Some points are of immediate interest. Firstly, there is a definite maximum free face area when operating with a domestic chimney of feasible diameter; this is a fairly modest half a square metre, and is relatively independent of chimney height. In practice however, taller chimneys give more reliable performance. Secondly, there is a definite optimum flue diameter and height for a given free face area. A small flue, 150mm diametre, can only clear an area of 0.175 square metre (0.42x0.42m) even with 6kW of heat being released to the flue. At 2kW a 0.5x0.5m opening would require a flue of 275mm diameter if 2m high, reducing to 225mm if the flue were 6m high.
This suggests that the simplest route to overcome smoke emission from large face areas is merely to increase chimney diameter. An extreme approach would be to build all flues 300mm or more in diameter; unfortunately, quite apart from constructional difficulties, large diameter chimneys allow such high air flows that there is over chilling of the flue gases, consequential loss of draught, and generally very poor operation. If exterior temperatures are higher than those inside the room, the chimney can then suffer "negative draught", and discharge flue gases to room! Figure 5 plots draught against flue diameter for 2 and 6kW at 2 and 6 metres height.
A minimum draught of 0.5 mm Water Gauge (WG) is recommended to prevent spillage into the room.
In order to appreciate the effect of wind on chimney performance it may be useful to note the velocity inside a chimney is typically equivalent to a "slight breeze" Beaufort scale 2: thus the slightly stronger but still common "gentle breeze" (Beaufort scale 3, and mean velocity 5 m/s) can have a significant effect on a short chimney in a downdraught situation. This latter has a velocity head of 0.13mm WG which is nearly one third of the suggested minimum design draught of 0.5mm WG.
As said above, the situation is compounded by short chimneys, thus a 2m flue of 0.20m diameter operating at 6kW generates a draft of 0.43mm WG. An equivalent chimney 4m high would generate 0.78mm. Where there is only a modest difference in theoretical free area a 2m chimney may be sufficient.
Chimney design procedure
Because of the complex interaction between chimney diameter, height and heat release; the design of flue systems is an interactive procedure. It is convenient to employ a computer programme of the type available from the author of this paper. The following approach is recommended:
1) Determine free face area of appliance and chimney.
2) Choose mean face velocity, typically 0.3m/s. Because of the velocity profile that occurs around any chimney inlet this can only be a mean value. Also the height of the opening (dimension 'h' in figure 1) should not be excessive (ideally less than 0.7m).
3) Determine chimney height. This should always be as tall as possible consistent with economic, aesthetic and engineering considerations. Ideally it will terminate above the maximum roofline of the
property. A chimney can never be too tall, excessive draught can be easily addressed, with a throat restrictor, or flue draught stabiliser.
4) Determine heat release to chimney under typical operating conditions. It may be noted from figure 3 that a greater heat release to chimney actually produces a greater chimney flow rate. The possibly unexpected result of this is that for an open appliance a large fire increases face velocity and hence reduces any tendency for smoke emission. It should be remembered that in an open appliance the volume of "combustion gas" (ie CO2), typically 10m /hr is very small compared with the total chimney flowrate of 300m /fur.
5) Determine suitable chimney diameter by means of above equations (given by programme). Chimneys should not usually exceed 300mm in diameter.
6) Check theoretical draught available within chimney (given by programme). This should be as large as possible and should exceed 0.5mm H2O. If this is the case the chimney should be satisfactory. All chimneys should contain as few bends as possible and preferably be straight.
There is a simple cross check available; from UK Building regulations (ref.7) the cross sectional area of the flue should not be less than 15% of the free face area of the fire or stove.
A three stone stove is to be mounted under a hood. This is to be closed in on three sides to leave one open side for access. This side will have a free face area, 400mm wide by 600rnm high; ie 0.24m². The maximum chimney height allowed on structural/cost grounds is 4m and typical heat release to chimney is 4kW. From the appropriate calculations and taking a face velocity of 0.3m/s the chimney should be 200mm in diameter. The theoretical draught is 0.61mm WG and chimney temperature should be 63°C (Amb 20°C).
If the free face area needs to be larger, say 0.6m wide by 0.75m high, the following comments would apply. Firstly the height of the free face area is greater than the recommended maximum, and secondly (from the appropriate calculations) even a 300mm diameter chimney is only just large enough. The draught would be only 0.35mm WG and flue temperature 44ºC. In practice such a flue would probably not be reliable and a 6m height would be preferable. If this is not viable, everything possible should be done to integrate the design of stove and hood to reduce the free face area. This might be with doors or screens.
1) A small free face area is necessary to eliminate smoke spillage into rooms. The maximum area should not exceed half a square metre.
2) Tall chimneys give more reliable performance.
3) The cross sectional area of a chimney must be no less than 15% of the free face area.
4) The average air velocity across the free face area should be minimum 0.2 m/s - 0.4 m/s to be sure that smoke will not spill over into the room due to minor air currents caused by movements or doors and windows.
1. Consultation on Indoor Air Pollution, World Health Organisation, Geneva, 5-8 June 1991.
2. I.V.H.E.Guide, London. Ventilation and Air Conditioning, Page B2-22.
3. JM.Coulson and J.FRichardson, Chemical Engineering, Vol 1, Pergamon, 1977.
4. BS 3376 - "Specification for Open Fires burning solid mineral fuels with convection with or without boilers": 1991.
5. Private Communication Mr D Wilkins, Coal Research Establishment.
6. Combustion Section Report No 182, March 1977, Coal Research Establishment.
7. UK Building Regulations. Department of the Environment, London.
Copies of the Chimney Design calculations (Copyright Coal Research Establishment) are available from the author at the Coal Research Establishment, Stoke Orchard, Cheltenham GL52 4RZ, UK, to whom all enquiries should be addressed.