Experimental analysis of swirl number and nozzle design for scale-up of partially cracked ammonia flames

Due to its ease of storage and existing global distribution network, interest in the use of renewably produced ammonia for decarbonising energy systems is growing. Partially cracking ammonia can overcome the flame stability challenges of this fuel, but demonstrations of high-power ammonia-based swirl flames with acceptable emissions have yet to be realised. Therefore, the present study examines the effects of varying swirl number and nozzle design on the static stability and emissions from 20 % (vol.) cracked ammonia swirl flames for a wide range of equivalence ratios (0.3 < Φ < 2.2) and thermal powers of 5, 10 and 15 kW. Additionally, a reference case of 100 kW thermal power at stoichiometric conditions was tested. Stable flames were shown across a broad range of equivalence ratios, swirl numbers and nozzle geometries although flame morphologies varied greatly. Of note was a geometric swirl number of 1.75 paired with a long nozzle, which enabled the transition to a flat, Coanda jet flow flame at equivalence ratios of 0.6 and 0.7. For a geometric swirl number of 1.45, shortening the nozzle resulted in significantly shorter, wider V-shape flames with greatly improved rich blowoff limits. This was found to be a desirable characteristic for reaching high thermal power with a constant nozzle throat diameter – i.e. dump plane velocity – as a widened flame brush prevents jet-like flames, which are susceptible to pinching off. This can also be achieved by increasing the swirl number, although to a lesser extent. However, with a widened flame brush, careful consideration must be given to confinement diameter to avoid flame impingement which has potential to increase local heat loss and hence reduce combustion efficiency, resulting in an increase in unburned NH₃ emissions. With the same geometric swirl number of 1.45, the shorter nozzle configuration resulted in higher NO emissions, potentially due to the shorter nozzle forming shorter, wider flames, meaning there was less residence time for NH₂ to consume NO in the flame zone. This difference was less noticeable at rich conditions, with all configurations reaching negligible NO emissions by Φ = 1.15