How to determine load cell response to arbitrary temperature changes with known step response(s)

We have a loadcell, the output of which is (naturally) temperature dependent. Under static load conditions, we have observed a dynamic response to temperature changes. Therefore, we did a step response experiment by placing the device in a -20 degree (C) freezer and then transfering it to a 40 degree (C) climate chamber.

What we see is that the temperature of the loadcell has a first order response and the loadcell output has a second degree response (under damped).

What we want to do now is to correct the temperature dynamics induced loadcell output error by using the temperature signal.

How can we use the results we obtained from the step response to correct the weight output when the temperature changes in real time, given temperature and loadcell output signals?

• Two measurement points would probably not be enough to get a good idea of how the system dynamics changes for other temperatures. You could use some sort of linear interpolation, but it could be a very bad approximation. Dec 10 '16 at 1:44
• You must at first prove that your system is linear before trying to derive an arbitrary response from unit response. This can be done by changing the temperature step size and then adding the second step time delayed. If the response observed would behave linear, then you can use the 1st derivative of the step response as the system transfer function and use the convolution operator to obtain response to the arbitrary input. Dec 11 '16 at 7:14