Model inversion

Model inversion theory is based on a Bayesian interpretation of the cause-and-effect relationship in a system, establishing the probability that a known response is caused by a follow-up of assumed sources. The inversion of models is born within the framework of geophysics, in an effort to model the interior of the earth from seismic data. The application of model inversion in various industrial cases allows us to respond to two common problems that usually arise, both scientific and practical.


On the one hand, the method allows the identification and characterization of different noise sources and their contribution to each point of the acoustic field, thus allowing interventions to be carried out on the primary sources to reduce it. On the other hand, it is a non-intrusive technique that allows minimal interference in the operation of the studied plant and in a very reasonable period of time.


ICR has developed proprietary computational tools for the study of the inverse problem that include optimization, resampling and stability analysis techniques to ensure the validity of the solutions obtained.


A case of successful application of the technology involved developing a methodology based on the inverse problem for a research project contracted by Ferrari Auto. ICR was able to determine satisfactorily the contribution of the different elements to interior cabin noise in the mid and high frequency range. The ICR proposal allowed Ferrari Auto to reduce the process of obtaining the acoustic power of the interior surfaces of the cabin from 30 to 2-3 days.