I’m working on a project to produce energy from ocean waves and wish to optimize the output of a small electrical generator that produces power (which I will measure) as several mechanical arms bounce up and down in the water. The machine floats on the water and the mechanical arms are driven up by a buoyant force as waves approach. I wish to do either a 4 or 5 factor DOE. The factors are: number of arms (3 or 6), length of mechanical arm (3 feet or 6 feet), buoyant float length (1 foot or 3 foot), central hub inertia (high or low), and arm counterweight (high or low). I have a single machine but the set up of the machine can be changed quickly into any state (about 2 minutes). There are too many interaction and the design is beyond simulations at this point. The machine will be taken into the ocean about 50 yards offshore on a sufficiently rough day and all trials will be run (I hope, depending on how many I have to do). My question is this: can I even use the DOE methodology? And if so, how? Should I run two or three trials of each treatment? And when a particular treatment is setup, should I repeat the data collect process about 3 times. and come up with and average reading? And the other question is, should I repeat the whole experiment on a different day when the ocean condition change? Or, is the DOE used only for processes that are under control and not for a mechanical machine that’s bouncing randomly up and down? Or do I have to try and construct several of the same machines to make any sense of the data?
Good afternoon,
I hope my answer is helpful, but I can't be sure it's a very good one ... but you can see it as a recommendation.
As the general procedure of the DOE methodology, it can be used in the following way: It identifies the input variables that can be controlled, also called factors. The limits or levels are chosen for each factor. Usually, two levels will be used for each variable. Note the upper level with +, the lower with -.
The design of the experiments is established and documented. This includes all the different combinations of levels (called arrangements or "treatments") that illustrate what levels of the variables will be for different combinations; how many times each arrangement will be subjected to the experiment (an action called replication); the sequence of all attempts, chosen in a random order (called randomization).
Regarding the number of studies for each treatment, I don't know if you have to do 2 or 3, but how much do you consider necessary to obtain a satisfactory result.
The data collection process should be repeated depending on how many studies you get considering the fact that in each experimental test the value of all the factors involved in the researched system changes.
It might be a good idea to do this experiment on another day when the state of the ocean changes, to make a comparison, just as a suggestion. Good luck!
All the best,
Goodbye!
- Badges
- Science topic
- Similar topics
- Force