Interpretation
Modélisation et identification d’un processus sidérurgique

In this way, the classic procedure using a suite of appropriate software, we obtain a number of domains corresponding to various tests. Model analysis can then begin. Do different tests with different frequency contents, carried out under different circumstances and several days apart, give the same result? Clearly, if we are content with local identification, we cannot, in non-academic cases, find the same point in the minimal criterion parameter space. Global identification will enable us to find an intersection between the various iso-D domains. If it's non-zero, we'll be able to determine a valid model for the various cases encountered. If it is zero, either the proposed structural model will be accepted but with a worse C ₀ level of adequacy so that an intersection appears, or the characterization assumptions will have to be repeated if...