Built in 2018 fall, the first year in my engineering study, the quadcopter "Donut" was equipped with PID control for stabilization and a GPS module for long range cruising and route planning. It could also be controlled with an RC controller.
Why a new design?
There is an old high-pressure boiler at the fluid instability lab, which was fixed by a traditional flange design. However, this also made it very difficult to remove the cap and gain access to the inside of the vessel for cleaning purposes.
The original boiler
The flange and the bolts & nuts
We noticed that they produced an interesting design called "Chain Engaged Fittings" (CEF) that could replace the traditional flange for high-pressure pipelines.
The design must be backward compatible so that it could be mounted on the original high-pressure vessel.
We would like the boiler to withstand at least 300 bar pressure.
There must be a way to quickly assemble/disassemble the boiler for easy access to the inside of the boiler.
A window is required for lighting purposes when performing high-speed recordings.
I planned to design a new hatch that utilizes the CEF design from Senter. To fix it on the original boiler, which relies on a flange to hold the cap in place, an adaptor is needed to connect the vessel and the hatch. For the lighting, I opted for an acrylic window that could be easily manufactured and replaced when broken.
Acrylic window - diameter
The opening from where we are able to look through should fulfill the lighting purposes. The dimension was decided by shooting a 10000 fps video while providing lighting through an opening on the original boiler. The opening was taped up to fix the diameter of the opening. This testing allowed me to determine the diameter for the opening on the hatch, and also provided some guidelines for the diameter of the acrylic window.
Acrylic window - Thickness
It is crucial that the window can withstand the 300 bar pressure. Luckily, a technical report from the Naval Civil Engineering Laboratory (Port Hueneme, Calif.) presents the exact information that we needed.
. D. Stachiw, C. United States. Naval Facilities Engineering, and L. Naval Civil Engineering, Windows for external or internal hydrostatic pressure vessels. Port Hueneme, Calif. :: U.S. Naval Civil Engineering Laboratory ; , 1967.
From the technical report, I was able to calculate the dimension for the acrylic window, which was based on dimensionless parameters calculated from the diameters of the opening.
The main reason we needed an adapter was to fix the hatch on top of the vessel. There were two variables that were free to decide: the length of the neck and the internal diameter. The internal diameter must be decided based on Static Analysis performed in CAD software, while the length of the neck was concerned about the interference: we need enough spacing for a torque wrench to fit in and screw the bolts and nuts.
A short animation was made to make sure that there is no interference and that the adapter could be easily be connected to the vessel body using the flange.
This part was designed for frequent assembly and disassembly, and therefore it had to be lightweight. The opening of the hatch was decided based on the lighting testing, while the thickness of the hatch was dictated based on Static analysis in Inventor.
The hatch also has to house a part of the acrylic window.
For equipment that has to undergo high pressure, sealing solutions are significant problems. At first, I was thinking about using O-rings to seal off every gap. However, some difficulties arose producing the grooves that I designed for the O-rings, and therefore I had to seek other ways to achieve the same effect.
Several options were proposed, and after discussing the subject matter with Senter Ltd and my advisor, I eventually decided to get rid of the O-ring design and instead utilize a pad produced by ourselves. Initially cut by hand using a razor blade, I eventually laser cut the silicone pad for more accurate dimension.
Manufacturing the acrylic window
I went to the school mechanics and specify the dimensions for the window using a lathe. Unfortunately, the first window was a failure due to a couple of reasons:
The rotation speed for the spindle was too fast, causing streaks of melted materials on a ring on the two flat faces of the cylinder, which made polishing difficult.
When cutting the material, some part of the material fractured, and although it did not seem like a big deal at first, the cracks quickly propagated and posed major concerns about its durability especially under pressure.
The diameter of the cylinder was machined from both ends, and at the place where the two passes met, there was a clear difference in dimension.
I realized that I did not specify my needs as clearly as I should have. The next time I went to the mechanic, I expressed my concerns over the first version and asked about possible solutions. This time, Mr. Huang, the skilled mechanic, took his time in tweaking the parameters to produce the best product. I observed how he tested several rotational speeds to make sure the material does not melt. Another interesting technique was that he utilized pressurized air to blow on the lathe. This could not only blow away the chips but also cool down the surface to prevent overheating.
Testing the boiler
Before being used in liquid nitrogen experiments as planned, we tested the boiler with a manual pump up to 200 bar, which was what we needed at that time. The structure held up as expected, and now the hatch could be easily removed after cycles of pressurization and depressurization. This allowed the operator to clean the inside of the boiler without breaking a sweat: simply loosen the bolts on the chain, remove the chain, and take off the hatch and the acrylic window inserted into the adapter. The adapter does not need to be removed, as the opening provided access to the inside of the vessel.
This design was really a game changer. Previously, it took at least two personnel, a torque wrench, and a painful amount of elbow grease to tighten the bolts to spec. Worse, there are 16 bolts, and it normally required around 30 minutes to tighten them all. It was also quite dangerous, too, as the cylindrical vessel made it very difficult to fix the vessel while applying torque onto the bolts.
Now, there is much less hassle to do the job. It takes less than 3 minutes to remove the hatch and the acrylic window, and the process is effortless.