Brainstorm: How would you run an experiment that simulates corrosion?
Brief for How would you run an experiment that simulates corrosion?
Assuming we have a solution that claims to detect corrosion under insulation, how would we test it? In other words, what are some creative ideas to "simulate" corrosion on a pipe?
Currently the best way to simulate corrosion is to etch away the surface of the pipe to simulate material loss. However, that is often far from the reality. An ideal would be a way to get a pipe to corrode within a few hours or days in a controllable manner. What is your idea for an experiment?
In response to a very good point raised by one of our members, here are some assumption to narrow the problem a bit.
1. Let us focus, for now, on an actual, controllable and measurable experiment that can work within a space of say 1 week. Let us assume (dream) that we have infinite resources. We can work on that issue at the next stage.
2. Even more specifically, let us assume we are working with an austenitic Steel pipe (pick any you want from the API list) of say 12inch OD and 12mm thickness. This is reasonably typical.
3. We would like to "generate" a corrosion causing a material loss of say 0.2mm per day or 1.4mm per week -- whichever you want to do .. resulting in around 12% material loss.
4. For phase 1 let us assume that the loss only has to be axially locatable -which means that I am only interested in where along the LENGTH of the pipe, the loss happens. We don't care where it is on the circumference.
5. The loss mechanism should be controllable -- meaning we should be able to STOP the corrosion when we have reached a particular point.
6. While we really want to simulate both ID and OD (inner and outer dia) corrosion, let us for simplicity, focus on ID for now.
7. Let us try and focus on generating the corrosion as opposed to the way we are testing it. We don't want to create a "defect" that can only be detected well by one method since the idea is to be able to compare the efficacy of multiple different methods.