I started the week by making a very fast draft of next year’s ROV frame, so that I could use it in a meeting with Martin Luther Culpepper. He is a professor of mechanical engineering at MIT.
We went through the design and calculated the expected forces across the tube and frame. We determined that the lateral tube strength was absolutely fine, having a 10x safety factor.
The strength in the other direction was less satisfactory. It only had a safety factor of 3x. Although that this is usually acceptable, because of my high pressure design inexperience, we decided that it would be beneficial to raise the tube thickness from 5/16 in to 3/8 in.
Another thing we discussed was the gasket and sealing mechanism.
He explained that a gasket would be superior to a O-Ring because of the high pressures involved in the system. He then gave me a few things to research to properly decide the number of bolts, what type of bolts, and the gasket I choose.
After showing him the flange-clamp design, he explained that the tolerance for the OD of the interior support portion of the end cap was crucial. He explained that the only way to find this tolerance is discrete element analysis. We didn’t have time to run through the problem then, so he offered to help me with it some time this summer.
I began researching gaskets.
I found PTFE, Viton, and Fluorosilicone to be the best options. Because the flange was not sealed to the tube, I needed a simple ring gasket with no bolt holes, that rested only along the rim of the tube. Because of the abnormal flange arrangement, no off the shelf components would work for what I needed.
I looked into custom ordering gaskets, however it would have cost in the range of 1 to 3 thousand dollars. So after researching the machinability of my options, I decided to use PTFE, commonly known as Teflon.
I found a supplier who sells 6.5 in diameter, 6 in long rods of PTFE. This would cost around 200$ before shipping or other fees. I estimate that it will in total cost around 300$.
After I decided on the gasket material, I looked into the design of the parts surrounding it.
In my meeting with Martin, he explained the concept of a force cone. This is a way of visualizing the strain put on an object by a bolt. He explained that these cones would need to overlap to cover the gasket to ensure a proper seal. After looking into the math for these force cones, I decided that it was too advanced for me to do in the moment, so I had AI make a python program to calculate the number of bolts required. This used a few factors including gasket surface area, gasket compression force, maximum bolt force, and bolt effective radius. I don’t believe that this is enough information, so I will work today to fix this program so that it gives me good data.
