I have installed accelerometers on a bridge on which train goes.
I don't have any force sensors on the bridge.
Now in this condition can I calculate the FRF of the bridge.
FRF=Acceleration function of bridge/force function of the bridge.
I already answered your connected question (https://www.researchgate.net/post/Can_I_make_a_force_sensor_from_an_acclelerometer_sensor), related to inferring force from accelerations.
From my simple point of view, you are looking for a transfer function of the system, that is a non-linear highly complex system. In theory you should apply a KNOW FORCE, possibly an IMPULSE and try to extract the transfer function from that. However, I suggest you to talk to someone that has an expertise in building structures, rather than looking at "sensor expertise".
Kind regards,
Daniel.
I think you cannot obtain a reliable transfer function from one sensor only. You can, of course, infer somehow the force, but the quality of the FRF so obtained will be the same as your inference. Knowing the structure (using an existing model) to the point of being able to compute the response when the train goes by (requires sophisticated software) and comparing it with the experimentally measured response is a way of checking the quality of the model and the software, but does not allow you to obtain any model from scratch. Another approach would be to record several travels, but you would need to be quite sure that all the relevant conditions (the train, its load, its speed, etc.) would be the same; then you just change the position of the sensor and would be able to obtain a transmissibility function (not the classical load-response FRF, but still a transfer function).
just my 2p...
Thanks you sir.Ok them kindly tell me what is the standard way of calculating the FRF of a bridge on which trains move.
I have many accelerometers with me.But no force sensor.
Measure the frequency of the acceleration signals after the train passed, when the bridge is in free movement without loads from train
FRF is the maximal responses of a structure under a harmonic force with various frequencies. So, is it suitable to say there is a FRF of a bridge under a moving train? In my research about high speed railway bridge, different frequency of external force could mean different speed of train. If it is just a train generally passing the bridge with a particular speed, you will never get a FRF of a bridge under moving train.
Please see if this helps.(. based on assumption that a linear system
is an acceptable model of the problem. validity to be established by you..)
1. Model the train as a moving gate function = mg *[u(x1- vt ) -u(x2-vt) ]
..x2-x1 =length of train . m= mass /unit length v = speed of train
2 Model the bridge as a filter H(z) = h0+h1*Z^-1 +...hn*z^-N (N=length of filter )
3 Obtain response Out(z) say of one train with known data . This is test input function
and experimental out put function .
4. Treat problem as System identification first . Use LMS or RLS
adaptive controller techniques to identify H(z) above
5. Now you can use it in reverse for any other train as you know H(z)
and Out(z) ..the output..
Cheers
6. If you have multiple sensors MIMO model may be tried.
7. Assume length of train less than span of bridge.
8 Divide bridge length into m parts where sensors are placed
9 Let U(z) = U1(z) , U2(z)...Um(z) represent sections on bridge...
the Ui(z) = U1(z) *z^-r ie each input is delayed by r (to be obtained from velocity
of train)
10 Let the bridge be [H(z) } m,m matrix representing coupled system). Here a
a suitable model for matrix eg circulant or Toeplitz for simplification can be considered.
!1 Use LMS or RLS adaptive system identification and proceed as mentioned earlier.
If we set m=1 we get SISO, could be the next step.. m=2 .
12 Another possibility is to use learning techniques like ANN.
Cheers
Good question.. Please share me the best answer might you trust...
Regards…
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training