Yesterday, three of us, David, Shawn and Sascha, worked furiously to do a test run of the oxidizer system with liquid CO2 before the evening ran out. We had some success and some lessons to learn.
The intention of the test was not only pressure test the system at similar pressures to the oxidizer we’ll be using and observe the behavior of the oxidizer injector nozzles but to also discover “unknown unknowns” before going into the full live tests.
Prior to setting up for that evening’s main CO2 event, we calibrated the emergency pressure relief valve using water pressure up to 1000psi. A few tweaks made by Shawn and the valve was good to go. Then it was on to mounting the oxidizer system vertically for the test.
Weighing Our Options
Partially filling the “run” tank from the main CO2 cylinder took longer than we expected. After 15 minutes we had only transferred 11 pounds of the original targeted 20 pounds – a fact we didn’t learn of until later (lesson #1: weighing of tanks before and after is critical for actual test to know how much oxidizer has been transferred and/or potentially lost).
Upon filling the run tank we noticed a solenoid valve was installed in reverse, and it appeared that it allowed flow even when not activated (that’s lesson #2).
As the CO2 slowly filled up, not surprisingly, the piping began to frost over. As more and more of the CO2 entered in as liquid and less as gas, the whole system began to warm up. We then saw the run tank pressure gently increase its pressure. Once it got to 500 psi, our low end for nominal pressure conditions in an actual live burn test, we decided it was time to pull the main flow valve and observe (albeit remotely).
Gone in a Cloud of CO2
In three seconds it was over. And in less than 1 second the video camera filming the output of the injector heads was completed enshrouded in a white CO2 cloud. We were able to barely observe the intended motion of the CO2 release but because the CO2 so quickly vaporized from liquid to gas it was unclear how well our injector head would atomize the oxidizer. This result was exactly the one Shawn had predicted, because the output is at atmospheric pressure and not a running combustion pressure of 500psi+ we just can’t observe how well the liquid is atomized by the injection bulkhead.
So while we could not observe clearly the performance of the injector nozzles we did learn some valuable lessons in both the setup of the solenoid valves, timing for pressurizing, procedural test steps and a “feel” for the behavior of the system under conditions similar to the roaring live tests we hope to initiate in upcoming weeks.
What’s next? Next up we begin incorporating the control and measurement software for the activation of the solenoids. And soon after we plan to do a series of short ignition event tests.