Impact force measured using an accelerometer mounted on the bullet - urgent please

[StathPol]

Hi guys,

I am performing impact tests on concrete slab. A bullet hits the concrete slab from 0.5 m. I am getting a graph that seems that the impact resonates at the head of the hemishperical bullet. I was expecting a pulse peak only.

I have syncrhonised the accelerometer with the video, however the graph is very weird, i think. is it possible to let me get your interpretation please?

[Ben]

How do you know it is not the wall oscilating?

Ben

[Matthew_Williams]

Could you be seeing reflections of the initial impulse?  If your time scale is mS, that might be about right.  If you calculate the speed of sound in your material vs. the material thickness, do you get a time similar to the period of your major oscillation?

I like the wall oscillation idea too; probably a combination of factors.

Matt

[StathPol]

Thank you guys for your responses.

No, i don't see reflections. The concrete's speed of sound is 3200 m/s and the thickness is 100 mm =0.1 m.

The concrete beam needs to rotate so I have a system with a rotation shaft; it does rotate but i do not know why i cannot get the one peak and then small oscillations. 😞

I have validated all of the sensors and the accelerometer with the high speed camera. SO, LabVIEW and camera are not a problem. Basically, I have double differentated the position-time of the bullet and then got acceleration.

The time scale in the graph is in seconds.

[Ben]

Thinking about it a bit more...

Some of those "follow on" peaks exceed the amplitude of the initial impact. That energy has to come from somewhere so I will specualte this morning that we are seeing the energy released as the concrete is cracking.

Is your camera/acquisition system fast enough to be abel corelate the carck propogation with the peaks shown on the graph?

Ben

[StathPol]

The DAQ samples at 100 kHz and the high speed camera samples at 10 kHz. Both are syncrhonised in DIAdem based on interpolation of the time.

I have done one more test this morning. I removed the clamping from the top surface of concrete to see what is happening.

The red graph shows the acquired signal.

The green graph shows the smooth signal of the raw acquired signal (red graph). I also inverted the red graph multiplying with -1 just in case the signals are wrong.

Also, I have focused near the concrete cracking.

[Ben]

From that last set of images it looks like the crack starts at the time of the first "mystery peak".

I would not be suprised if those other oscillations that you have filtered out may be the reflections of the energy bouncing around in the medium. (I know you looke at that previously but did you analyze the smaller amplitude singal that you are filtering out?).

Not tha tI know what I am talking about.

Ben

[Alexander_Sobolev]

From a quick search (http://www.diva-portal.org/smash/get/diva2:563604/fulltext01.pdf) concrete has a very!!! nonlinear dependence of dynamic modulus (and hence speed of sound) vs frequency. It has multiple peaks.

This means 

1) narrow peaks travel awfully 

2) if you place accelerometer close to impact point, your graph will be different

3) Resonant frequencies will shift depending on block geometry. 

4) Can you do FFT of your initial graph (not filtered), will you get something as above? There definitely are separate frequencies

So I would say that you have bouncing of block as a whole. High frequencies (that you filter between red and green graph) start when cracking starts, they can be related to grain size and crack speed. Simple impact has low energy in these frequencies and you do not see them before the crack starts.

[StathPol]

I went back to the electronics room to simulate the test in the concrete. I just dropped the bullet from a small distance, and the graph looks as expected. 

The actual test is as shown below. I am aware that the bullet nose plays an inpotant role on the impact response. In the actual test, we use a steel rod connected with the bullet to add incremental weights. The front face of the bullet is not round, however I am not quite sure if this affects the impact response.

[StathPol]

The rubber pad on the top surface of concrete was not used, so please ignore it in the image.