Dear Mr Chong,

EMA and OMA are procedures where we fit an assumed model to a measured result.

In doing so, we try and solve an inverse problem. To exemplify, an inverse problem of the simplest kind is that if I tell you that the answer is four - now, please tell me what is the question. For obvious reasons, the problem posed has an infinite solution set.

Similarly, the measured data that serves as in EMA/OMA can be fit using a multitude of models. This fit can range from very good to garbage and it is up to the analyst to decide whether a model fit is any good.

The damping parameter is the weakest variable in this fit as it affects response only at resonance, it is based on simplistic theory and is based on the assumption of being evenly distributed. Hence, it tends to associate with large uncertainty.

Sincerely

Claes