C

#### C McK

I am trying to develop a closed loop feedback control system for a whisky filling machine using compressive load cells. However I can't find a transfer function for the load cell. Can anyone help?

D

#### Dr ir Gerrit M. van der Molen

The simplest typical transfer function for a load cell would be:

y/u = Kg*(K1*s + w^2)/(s^2 + 2*Kd*w*s + w^2)

Where:
s = Laplace operator 's'
y = load cell output signal (voltage or current?)
u = weight on load cell

Parameters:
w = natural frequency
Kd = damping factor
Kg = overall gain
K1 = 'other parameter' (you can probably assume that K1=0; I am not sure how you would determine this if you are working from manufacturer's data
rather than test measurement data)

To determine the parameters:
Kg: this is the ratio of measured output (voltage, etc.) to weight, which you can probably get quite easily from your data.

w: this is the frequency (normally expressed in rad/s) at which the load cell oscillates when a weight is (suddenly) put on it. This may be in,
or derivable from, the manufacturer's data.

Kd: the 'rate' at which the above oscillations decline (0=no decline, 1='critical damping', i.e. no overshoot; usually Kd is approximately 0.5
to 0.7)

As Jim noted in his earlier reply, there are other factors (non-linearities) that significantly affect the actual behaviour of the
load cell, compared to the idealised linear behaviour of a transfer function. Keep these in mind!

Hope this helps,

Gerrit.

-----------------------------------------------------------------------
Dr ir Gerrit M. van der Molen
Industrial Systems and Control Ltd.
50 George Street, Glasgow G1 1QE, UK
Tel: (+44) (0)141-553 1111 Fax: (+44) (0)141-553 1232
Email: [email protected] WWW: http://www.isc-ltd.com
-----------------------------------------------------------------------

J

#### James Fountas

Not having done transfer functions for a while, I would think that the transfer function for the load cell would be trivial. I would assume the difficulty would be in the modeling of fluid momentum, force applied to the load cell, and actual volume discharge as a function of time. These other factors would be based on process equipment design.

Using transfer functions for the basis of control wouldn’t be my first choice.

James Fountas
[email protected]

G

#### Gilles Habel

Schneider Electric offer a card (TSX ISP Y100)that accept directly load cells (up to 8 with a junction box)and convert the weight directly inside a memory location of the Premium PLC. In the instruction set of this PLC, you have closed loops instructions to implement process control. You could also have the weight display localy with a dedicated "Plug and Play" display.

Hope this will help,

Gilles Habel