Can chemical process industry benefit from analog electronics diagnosis methods?

The present paper will not answer the question asked in its title! It would require a deep knowledge of both domains (chemical processes and analog circuits) that we don't have. Nonetheless, we have the feeling that the method we recently developed for diagnosing analog electronic circuits is far more general that its initial purpose. This confidence is mainly based on the fact that our method not only offers an original way to take into account differential equations -which are obviously at the basis of process modeling- but also to consider inaccuracies such as measurement errors, parameter dispersion, which are often an impediment when correctness of diagnosis is concerned. Therefore, the paper is limited to a description of our method, illustrated on very simple electronic components such as resistors, capacitors or inductors. The analogy with pipes, tanks... is obvious. Its application to process diagnosis is less obvious since many other aspects than component modeling must be taken into account, for instance: what is to be diagnosed, what ambiguity is acceptable, what are the measurements... Some of these points are also discussed. Copyright (C) 1998 IFAC.