Process control paradigm ~ Process Control for Dummies

I started to learn process control in the undergrad course and has been fascinated by it ever since. Now, I am in a grad school studying more about process control focusing in Model-Based Control (MBC). What is MBC?? Well, we will cover the topic later on. For the introduction, process control can be divided by three most major categories which is Modelling, Control and Optimization (refer figure below).


Figure 1 Process control overview

The modeling is the part where we developed model which is a representation of a real system or process. Why we need a model? Well, for some reasons there are some limitation using a real system or process e.g. operation cost, safety issues, equipment constraint, etc. Thus, if we can develop a good representation of the system/process, then we can use the model to represent the actual one. After we get a good model, then we need to figure out its control system. Why we need to control? Usually, a process/system need to have a mechanism to help it to maintain its operation as some 'disturbance' can happen. These 'disturbance' can affect the system/process operation so it would not achieve its target. So, some sort of mechanism is needed to reject the 'disturbance' and to properly shift the operation level if needed. Generally there are a wide selection of control strategy that can be used. After a controller had been set up, the final part would be the optimization. Optimization study is needed to further improve the process/system capability such as to maximize its operation output and at the same time minimize its operation cost. This study will help in term of economic assessment of the system/process e.g. minimize energy consumption, minimize process duration, maximize conversion etc. In a way, modeling, control and optimization may act in a hierarchy order. However, there are people that just focus on a single subject, some may pair it up and some just do all of them at once.

In a nutshell: Process control is a study of how we want to control something in a proper way. In addition, controlling using a model is a convenient way to design and test a controller system before implemented it in the real process/system. After control, it would be much better if we can optimize the process/system somehow to get a better deal of it.