As one looks to optimize, troubleshoot or simply better understand your plant operations, you need a model. An even more crucial question is does the model represent the plant? Building models is fun, and everyone likes it. But when confronted with how well does it match with the plant; it is a much more involved question. There are various ways of matching the plant with the model, and we describe in this blog two different ways of matching the plant data with the model. Whichever approach one takes depends on the context, and the ability to achieve our goals in a sustained manner all the time.
While it is context dependent, we generally prefer a first principles based equation-oriented modeling approach to managing a model and matching it to plant data. And we suggest to complement this with second-order or equivalent reduced order hybrid models.
In the attached video you will see more details on two most common approached.