Cheap acceleration metering? / G Logger
September 2, 2008 at 4:34 pm #683
I know some phones and other things are starting to have acceleration meters inside. I would like to be able to measure accelerations on the models I am testing. I think what I want is something that will record things and then let me download into my computer over USB. Does anyone have any experience or recommendations?
Googling I found:some. One that looks good. A 100 Hz sampling rate is plenty fast enough. It is waterproof. And cheap enough.September 3, 2008 at 9:39 pm #3719
That data logger thing looks convenient, and it isnt insanely expensive.
If you happen to own a wii: a friend of mine did an EE project where he used wii controllers to locate robots: its supposedly quite easy to get going and read out its data, including the accelerometer, but setting up a wii in your kayak is probably not very alluring.September 3, 2008 at 10:02 pm #3720
Just a word of warning about tilt sensors and rate gyros — they are very noisy. It takes a fair amount of post processing to get useful data out of them. In the robotics field, we are almost always forced to come up with a Kalman filter to try and reduce the data down to something usable.
One of the tricks the wave tank people use is that they put 3 LED’s in a triangle on top of their platform. Then they take a video of the resultant motion from two views. Next they digitize the LED locations. Using stereoscopic rules, it is possible to compute the position of the platform in 3 dimensions vs. time. From that it is possible to compute all of the velocities and accelerations. This strategy will not work well in a kayak.September 3, 2008 at 11:45 pm #3721
Since the main acceleration is probably close to a sine wave, at least when waves are mostly coming from one direction, filtering the data is probably going to work out ok.
I might be able to have a tripod on the pier and have the little island or the other side of he harbor in the background. Then using the width of the model for scale I could measure how much it is moving up/down and right/left over what time. Since the main waves will be coming from my right going to my left there is not much motion in the 3rd dimension, which I won’t be able to measure. But I could only do this with models which go less than 8 feet deep. I may even have some video I can do this with as we did do some tripod from the pier shots, though most were with the model upwind of the pier and only ocean in the background.
I am most curious about the accelerations on this ball with the 23 foot hanging ballast. Looking at some of those videos, the clouds are essentially still background for the couple seconds of a wave cycle, so they are as good as land for a background. The problem is just that the camera either with me in the water or on the kayak is not still. Might be a shot that seems still enough for the main motion of the ball so I could estimate things though. Hum. Really the ball goes up and down with about the amplitude and period of the wave. So I can estimate the up and down acceleration I think.September 4, 2008 at 1:08 am #3722
Since my ball model comes close to moving up and down with the wave, I am trying to calculate what sort of vertical accelerations a full scale ball on a normal wave would have. So let me call an 8 foot high wave with an 8 second period normal. Now in 8 seconds it has to accelerate up, slow to a stop, accelerate down, and slow to a stop. So lets look at just the first 1/4th, just a 2 second part of this. In 2 seconds of acceleration we cover 4 feet. From physics I remember s= 1/2 a t^2. so a = 2 * s / t^2 = 2 * 4 / 4 = 2 feet per second^2 Since 1 G is 32 feet per second^2, this is 1/16th G
I think the acceleration is the same the whole loop (particles of water do a circle).
Anyone see anything wrong with this math?
I have been on boats where we were doing 0 G and then 2 Gs. So +-1/16th G from normal does not sound bad.September 4, 2008 at 6:24 am #3723
You are assuming constant acceleration, and therefore getting too low a number. I think wed get a more accurate number by assuming harmonic motion, then:
position = 8 * sin(t/8*(2*pi))
Acceleration is the second derivative of position, so:
acceleration = -1/2*pi^2 * sin(t/8*(2*pi)), which has a maximum of ~5 feet per second^2
I dont know how bad that is, but moving with the waves is not very comfortable to most people over sustained periods.
And im not sure this scale model can be simply scaled up. The mass is very important: what mass / water displacement does your current model have? I think a heavier model that lies deeper in the water would tend to move much less.September 4, 2008 at 7:29 am #3724
>You are assuming constant acceleration, and therefore getting too low a number.
>I dont know how bad that is, but moving with the waves is not very comfortable to most people over sustained periods.
I am not sure how bad that is either. In my experience a raft or catamaran moving with the waves is fine but a monohull that has other tipping going on can be very uncomfortable. Also, a catamaran moving through the waves is much more bumpy that one just drifting with the waves. Really would be good to get more information about what accelerations are ok.
>And im not sure this scale model can be simply scaled up. The mass is very important: what mass / water
>displacement does your current model have? I think a heavier model that lies deeper in the water
>would tend to move much less.
Yes. I kind of waved my hands and said, the model follows simulated 50 foot waves so lets assume it follows 8 foot waves. But really it does not move up and down the full height of small waves. But I have not been where I can test my 23 foot hanging ballast when there were 4 inch waves there and have little hope of doing that really. My model has enough ballast that the ball is about half in the water and half out. If you do the same thing in a full scale version it will move like the model in scaled waves and scaled time. See this page for more on this:September 4, 2008 at 7:47 am #3725
There are studies and formulas for “motion sickness incidence”:September 4, 2008 at 4:17 pm #3730
There are 2 graphs in the page below. I think this is saying that with 0.1 G for 2.5 hours about 25% of the people will throw up. Seems to be still rising with time too.September 4, 2008 at 8:58 pm #3734
What just occured to me is that the wavelength relative to the dimensions of the structure are probably the most important variable.
If the waves are sufficiently big compared to the structure, the structure will tend to follow them perfectly. If the waves are microscopically small, the structure will not folow at all.
I think matching the relative wavelength is the most important in making a usefull comparison. So thats a good thing: 8 feet to a scale model is rediculously long to a real structure.: there is hardly any wave amplitude at that frequency probably. What is the wavelength of a typical stormy and calm weather wave, and how does that compare to the scale model?
I know about dimensional analysis: when i have the time to think about it i hope i can give more complete arguments.September 4, 2008 at 11:56 pm #3735
Yes, the ratio of the model wave size relative to the model should be the same as the full size wave to the full size structure. So for the 1:25 scale that I use a 4 inch wave on the model is like an 8 foot wave on the full scale structure.
A typical wave depends on where you are. In the Caribbean it might be 6 feet and in the Atlantic 8 or 10 maybe. Period seems to be 8 or 10 seconds too. A 10 second period on the full scale should be 2 seconds for the model. The worst storm waves can get up to 100 fee hight, but I think we can migrate around and avoid that situation. I think you could stay under 25 foot with this. So for models that is 1 foot high waves.
In here you can find graphs of “velocity vs period” and “wavelength vs period”: We are interested in deep water waves. Using that graph you could see that a 12 second period means a 250 meter wavelength.September 5, 2008 at 7:25 am #3736
Ah by 8 feet you meant amplitude.
I was talking about wavelength, not amplitude.
I made a mistake by the way: it should be 4 * sin, etc…
Amplitude of water waves is defined as crest to trough height, so that yields 2.5, which is a lot closer to your figure of 2.September 5, 2008 at 9:28 am #3737
I think we need to understand what is acceptable motion for a seastead design. I started a wiki page for that.
It sort of looks like human brains can adapt and that after a few hours or days you get your “sea legs” and will be ok. If tourists are coming for 1 week visits, a few days to adapt takes away a lot of the fun. If a family is moving onto a seastead for good, a few days to adapt is no big deal.September 5, 2008 at 11:35 am #3739
25% vomit ratio on a small seastead seems like an acceptable figure to me. The early small structures will be aimed at adventurous types anyway.
That graph looks to me like it levels out after 25% with 0,1G.September 5, 2008 at 2:47 pm #3740
>25% vomit ratio on a small seastead seems like an acceptable figure to me.
I think so too. If a small seastead is better than a live aboard boat in:
2) living space
Then I think there will be a market for it. It does not have to be perfect in any of these ways, just offer a better overall tradeoff for people who are not in a hurry to get anywhere. I think the stability of the Ball House is much better than any boat near its price range. So even if it moves up and down some, I don’t think that would be the main thing limiting sales.
The forum ‘Structure Designs’ is closed to new topics and replies.
Posted on at