February 2024
Igniter Hotfire Testing Campaign
Prior to 2/1/24 I had only hotfired my torch igniter design once in May 2023. The 650ms duration hotfire in May was on an old horizontal test stand I had been using for other projects. The kerosene was fed from an automotive booster pump and the GOX was fed from a Bernzomatic canister. The hotfire plume color was bright blue despite planning for an O/F ratio of ~1.80 which was the first hint that my orifice Cds were slightly off. When I decided to return to igniter development using the new test stand I put together a new, slightly modified version of the igniter with a new GOX inlet and a threaded stainless steel throat module that will eventually allow the igniter to be installed on the thrust chamber. GOX was fed the same way as the test from May except I started using a ball valve with a custom actuator as the main GOX valve instead of a direct acting solenoid which was less oxygen compatible. Instead of using the old automotive pump for kerosene, I connected the igniter to the main engine fuel pump. I also replaced the old igniter fuel solenoid with a higher pressure parker-skinner solenoid I found at a local industrial surplus store.
On 2/1/24 I successfully hotfired the new igniter iteration for a full duration test but the op was slightly a mess and I forgot to record video which felt terrible but I did log data. I recycled for a second test by which time the leftover GOX canister from May was too low and the chamber appeared to quench when the mixture ratio intermittently got too rich. I bought more GOX and ran another test on 2/3/24 which ran for full duration of 1.3 seconds. Video showed the plume to be a pale blue at which point I was pretty certain there was something skewing my calculations. I had been calculating Cd using an equation for sharp-edged orifices but the true geometry of the GOX orifice included a conical entrance and the L/D of the orifice was actually greater than 1. I ended up running 2D axisymmetric CFD in Ansys Fluent which suggested a significantly higher Cd for the GOX orifice. Additionally I inspected the extremely small fuel orifice under a microscope at my university and used imagej to measure the diameter and found that it was smaller than I thought. I implemented my findings in my code to plan new pressures for future tests. The higher Cd of the oxygen orifice was particularly significant since the higher flow rate would result in a higher chamber pressure, decreasing the DP across the fuel orifice, lowering its flow rate and strongly biasing the igniter towards a lean mixture ratio.
Based on the results of the failed test from 2/1/24 I was worried that too much liquid fuel injected in the chamber would cause stability problems for the igniter compounded by the already low residence time. For the subsequent test I aimed to vary the mixture ratio throughout the test by throttling the fuel pump pressure up and down in the autosequence. This would allow me to observe the effect of different mixture ratios on the apparent performance and stability of the igniter and asses the accuracy of my new calculations by viewing the changing plume color. It was important for me to verify that a low MR was possible since I am uncertain if it will be safe to run the igniter at a high MR when installed in my aluminum thrust chamber. On 2/11/24 I ran the MR sweep test which went for a full duration of 2.0 seconds. The original test profile had the igniter start at an O/F of 2.60, then after a short time the fuel pump would ramp up power, lowering the O/F to 2.10, then the igniter would return to an O/F of 2.60 and shut down. The initial state at 2.60 ended up mostly overlapping with the startup transience so the video shows the igniter starting at the low MR and increasing over time. The oxygen reg pressure was slightly lower than planned which made the final low and high MR setpoints closer to 2.00 and 2.45 instead of 2.10 and 2.60. Next steps for developing the igniter are to hotfire it while installed in the engine and potentially run many additional tests to demonstrate reliability and build a strong family of data. The original plan for the igniter was to run it off the LOX pump using a heat exchanger as well and this will need to be tested eventually however I am also considering just sticking with the GOX canister for simplicity and greater reliability due to higher and more predictable oxygen temperature.
