Wave Buoy data logger : Accelerometer/Gyroscope/Compass/Clock

Hi folks!

I've been given a project of a data logger. I'm looking for collecting and transferring raw sensor data via a GSM module. The sensors are : MPU6050+clock+magnetic compass (need to know where the North is for processing).

Once at sea, the unit has to be entirely autonomous : self-powered, collecting and transfering data independently. At that point of the project, I already have an Arduino board with a MPU6050 (Acccelero+Gyro) collecting data onto an SD card, powered with a 9V battery. I don't have the magnetic compass and the clock yet. The buoy is moored and designed to get/transfer data for at least 24 hours.

My concern is about the accuracy of such sensors. Indeed, we then have to process all these data to determine the wave period, wave height, wave direction, the spectrum...we need sensor on which we can rely.

According to a colleague (oceanographer), the sampling frequency doesn't have to exceed 4Hz and an 'inconsistent' sensor isn't a problem as long as the error doesn't change during the process (so we can remove the offset eventually). Another idea would be to combine 3 sensors to counterbalance and filter the data. Another engineer told use about the inaccuracy of the MPU6050's gyro, as it drifts far to much to get proper data when considering long wave length (swell,heave), this gyro mainly made for a quick response measurement. It's apparently not suited for that type of application and keeps on drifting and spinning, giving a wrong idea of the actual inclination.

The position of each instrument inside the canister/hull is also important. I was thinking to place them close to the metacentre/gravity center of the hull. In the best case scenario, it would reduce noise and unnecessary offset.

What would be your advice/solution to implement and place the right instrument on the right place? What is your opinion about combining sensors as a filter and the gyroscope drifting and the different assumptions made above ?

Many thanks in advance, Cheers!

umfana:
My concern is about the accuracy of such sensors. Indeed, we then have to process all these data to determine the wave period, wave height, wave direction, the spectrum...we need sensor on which we can rely.

Hi,

I've thinking a while on the issue. Not in -logical- order:

  1. Your goal is to get the "elongations" (distances from the buoy actual position to the "completely flat" sea surface one) spectrum; although a double integration on the acceleration value will give that, I doubt that in a long time period -elongation- data would be consistent. Maybe there are some mathematics -discrete data calculations, I mean- to achieve that.
  2. As you cant determine the -moving- buoy tilt, you cant determine vertical acceleration. Can the accelerometers be "bolted" to the buoy body or you need a kind of giroscope -to mount them?
  3. For such an imprecise measurement -because its own nature-, I dont think that sensor accuracy is a real problem. Nevertheless the idea of mounting several sensors is good to discard a wrong one.

Hope it helps.

Regards.

There are plenty of research papers on line relevant to your questions, for example this one and this one.

Time to do some research! One place to start is to google "wave buoy data logger", which incidentally turns up commercial offerings.

There is a drift buoy project with a buoy that has been at sea for about 700 days. A good use for Irdium satellite radios.

http://mdbuoyproject.wixsite.com/default/blog