Deformation measurment of metall reactor

[Fur.1]

Dear all,

​

I am interested in the MEMS accelerometer (IIS3DWB) to measure radial and maybe vertical movement of a metall reactor. The maximum g range of the movement should be 4-5 g. And the vibration frequency should be 0.5-2 kHz. I would install approximately 6-8 sensors at orthogonal positions on the reactor (3-4 meters apart).

​

Now I would like to ask you, if you could suggest a combination of components with which the accelerometer can be used. The goal would be to install the sensor combined with an evaluation board, or whatever you suggest, directly on the reactor. We would then need a way to transfer the raw data on a computer system. The data sampling frequency should be as high as possible (around 10kHz).

​

Thank you in advance for your help!

​

[Ozone]

As a sidenote, you could do a force measurment with strain gauges alternatively, or additionally.

> Now I would like to ask you, if you could suggest a combination of components with which the accelerometer can be used. The goal would be to install the sensor combined with an evaluation board, or whatever you suggest, directly on the reactor. We would then need a way to transfer the raw data on a computer system. The data sampling frequency should be as high as possible (around 10kHz).

I would suggest to think about data collection and thus bus connection first. With that, I mean connection from the local controller board to the main unit, perhaps a PC.

For one sensor with 3 axes a 16 bit data and 10kHz, you have about 0,5 Mbit/s requied bandwidth. Suppose you want equidistant sampling, you need a bus that sustains this bandwidth, and a protocol supporting two-way communication. In sum, 8 sensor units would produce about 4MBit/s.

I first thought of CAN, but standard CAN maxes out at 1Mbit. And CAN-FD seems a bit experimental currently.

AFAIK, no UART-based bus allows for such data rates with the required number of peers and physical distances.

Perhaps ethernet.

The data collection part from the sensor itself is (almost) trivial. I know of no ST eval board that does not support I2C and/or SPI, which the IIS3DWB requires.

[Fur.1]

Thank you Ozone for your detailed answer!

I will have a look whether ethernet would be possible or not.

We have already considered strain gauges and came to the conclusion that accelerometers would be better for our application.

[Eleon BORLINI]

Hi @Fur.1​ , as regards your question

>> Now I would like to ask you, if you could suggest a combination of components with which the accelerometer can be used. 

I would suggest you to use an appropriate number of STEVAL-MKI208V1K, which can be attached with a tape on your vibrating reactor and whose geometry is optimized for the vibration monitoring in industrial environments. You can wire a number of on these evaluation board to a standard X-Nucleo board, and manage the SPI communication this way. The standard C-drivers for the IIS3DWB are available on Github repository. Unfortunately there is not a ready-to-use plug&play solution for your application in ST catalog.

Regards

[Fur.1]

Thank you Eleon for your answer!

I will discuss this with my colleagues. It may be a really good solution.

Kind regards