The use of solid biomass stoves for heating: An investigation into the effects of fuel properties and operational practices on pollutant emissions

There is considerable public interest in the emission of pollutants, especially smoke, from wood burning stoves in the UK. The popularity of residential combustion appliances has increased in recent years which has contributed in poor urban air quality. Additionally, the use of inadequate fuels and limitations in stove testing methods may suggest that the negative impact of such appliances is worse than currently predicted.
Combustion testing was undertaken on a HETAS approved 5.7 kW Waterford Stanley Oisin SF NB multifuel heating stove. A custom LabVIEW development platform was constructed for measuring various combustion properties including burning rate, flue gas flowrate and temperature. Gaseous pollutant analysis was undertaken using fourier transform infrared (FTIR) spectroscopy and electrochemical sensors. Particulate pollutant sampling was undertaken using a Dekati PM10 Impactor system. Independent studies investigated changes in combustion conditions and pollutant formation associated with (i) changes in the properties of the fuel and (ii) the influence of the stove operator. Predominantly, the impact of fuelwood moisture content, cold-start operation and reproducibility where reviewed.
Moisture content (MC%) was shown to significantly affect the properties of the combustion reaction. The combustion of wet fuelwood resulted in lower combustion temperatures and a prolonged burning period (reduced burning rate). The emission of CO, CH4, non-methane volatile organic compound (NMVOC) and CH2O was shown to increase with (i) an increase in MC% and (ii) the subsequent reduction in combustion temperature. The average CO emission for dry fuelwood (MC is <20%) was 59.0±22.4 g/kgfuel and for wet fuelwood (MC is >20%) was 105.6±12.3 g/kgfuel. The average CH4 emission for dry fuelwood was 1.32±0.76 g/kgfuel and for wet fuelwood was 5.01.6±2.48 g/kgfuel. The average NMVOC emission for dry fuelwood was 2.14±1.26 g/kgfuel and for wet fuelwood was 8.51±4.17 g/kgfuel. The average CH2O emission for dry fuelwood was 0.47±0.36 g/kgfuel and for wet fuelwood was 2.00±1.09 g/kgfuel. Both NOx and SO2 emission was shown to be generally unaffected by MC% and was dependent upon fuel nitrogen and sulphur content. Particulate matter (PM) emission was also found to vary depending on MC%. The combustion of dry fuelwood produced lower PM emissions (2.03±1.06 g/kgfuel) than wet fuelwood (7.20±4.85 g/kgfuel). As a result, combusting fuelwood which maintains Ready to Burn certified moisture contents contributes in a PM emission reduction of 112%. In addition, the structure and morphology of soot generated by the combustion of dry and wet fuelwood was also found to differ. Dry fuelwood produced soot comprising long chains of spherical particulate matter which presented a higher elemental carbon (EC) fraction. Alternatively, wet fuelwood produced amorphous tar-like material which presented a higher organic carbon (OC) fraction.
The impact of repeat stove testing was examined in order to determine the effect of prolonged operational practices and emission inventory sample size on the properties of the combustion process and results confidence. An increase in the inventory size resulted in an improvement in results confidence to an extent. Issues relating to reproducibility were observed when additional batches were applied to the stove. An increase in combustion temperature was identified during prolonged stove operation which improved combustion efficiency and reduced reproducibility. Poor stove operation was also presented during prolonged operation in response to limitations in the fuel. Finally, individual events relating to the dynamic nature of the combustion reaction limited reproducibility. These events, such as a fuel particle falling from the heated grate, leads to an increase in emissions and is identified as an uncontrollable factor.
Finally, the inclusion of cold-start operation within standard operating practices and emission inventories was investigated. Cold-start testing is often negated from standardised testing procedures due to the difficulties in attaining repeatable results. The findings of this work revealed a notable increase in the average emission of CO by a factor of 2.9, total hydrocarbon (THC) by a factor of 2.4 and PM by a factor of 1 when cold-start data was included. Data from stove testing is applied in the regulation of appliances and in climate modelling. These results show that by neglecting cold-start data from accreditation and testing practices, emission inventories may be underestimating the true impact of stove operation. This is of greater importance in urban locations where air quality problems persist, and stove operational duration may be confined to a single cold-start batch only.