Development and performance characteristics of a family of gas-fired pulsed combustors
Two nominally 15, and 30 kW Helmholtz-type pulsed combustors were designed and constructed. These were bench mounted with the heat exchangers (i.e. combustion chamber and tailpipe) immersed in the water bath. Their design was based upon the design of a nominally 7.5 kW pulsed unit previously developed at Middlesex University. The design enabled the lengths of the combustion chambers to be varied so that various combustion chamber volumes could be achieved. This provided a new dimension to the study of pulsed combustors which is lacking in many reported works. It was found that the required input rates could be achieved by scaling up or down each combustion chamber dimensions linearly by a factor of 1.5, while maintaining the geometry identical. Tests showed that the present design of pulsed combustors can operate successfully at various input rates of mains natural gas (93 % methane) with a maximum turn-down ratio of 1.8:1. Results indicated that the three developed combustors would generally operate in the fuel-lean condition. Interestingly, these tests revealed that the amount of excess air reduced as the combustion chamber volume (CCV) was increased. Systematic investigation on the three developed combustors showed that the temperature within the combustor was principally controlled by the air-to-fuel ratio (A/F). Analysis of the average measured NOx concentrations at various operating conditions indicated that NOx emission in this type of pulsed combustor is principally controlled by combustion temperature with no significant influence of combustion chamber volume, tailpipe length or scale of the combustors except in so far as these influenced the A/F and hence the temperature within the combustor. The dominant role of temperature on NOx production from these combustors become more evident when nitrogen or argon was injected into the system resulting in reduced NOx emissions at a given A/F. Systematic analysis of data indicated that as the amount of diluent increased, the temperature within the combustor decreased. Almost all the NOx values recorded were in the form of NO which is believed to be as a result of the high flame temperature (typically above 1850K). The minimum recorded NOx value was 5 ppm at the upper limiting value of excess air ratio, λ ; importantly it was round that at these high A/F values there was no significant reduction in overall efficiency of the pulsed units, showing calculated values above 90%. Analysis of data indicated that combustion temperature is also a primary factor controlling CO emissions from the present design of pulsed combustors. CO concentrations exhibited U-shaped characteristics when plotted vs λ, showing maximum values at the lowest and highest λ values. By changing water bath temperature (WBT) and hence modifying heat losses to the combustion chamber wall, it was shown that the quenching of the combustion reactions and incomplete mixing of air and gas prior to combustion are contributing factors to CO formation in this type of pulsed combustor. The developed pulsed combustors were operated successfully with standard test gases. The composition and flame stability of these test gases were similar to the standard test gases G21 (incomplete combustion gas), G222 (light back gas) and G23 (flame lift gas). Analysis of the exhaust gas composition showed similar trends to those obtained when burning mains natural gas; as the heat input was increased, O2 levels decreased while CO2 and NOx emission levels increased. Similarly, CO concentrations showed U-shaped characteristics when plotted against firing rate. Measurements of peak pulsing pressure and frequency were used as a guide to operation and stability performance of the pulsed units. It was found that the operating frequency was a function of configuration of the combustors and temperature of the internal gases. Frequency of operation showed a reciprocal correlation with volume of combustion chamber and tailpipe length and increased as the heat input was increased. Pulsing pressure amplitude also was influenced by change of configuration of the combustors, increasing as the CCV and tailpipe length were decreased. Analysis of experimental data obtained at fixed configuration of the combustors showed that the peak pulsing pressure was a strong function of the heat release per cycle in the present design of pulsed combustors. A major drawback of the use of pulsating combustors is the high noise level which is associated with their operation. It was found that it is possible to reduce overall noise levels of the pulsed burners to acceptable values by configuring the system appropriately. This included the use of expansion chambers at the inlet and the exhaust outlet which reduced the overall noise levels to a minimum value of 65 dBA.