7 replies to this topic

[hoyoku]

Dear all,

I have a few questions to ask:

a)I would like to know under what circumstances, the product concentration from the outlet of the reactor would be an important variabel to be controlled.

b)Is it possible to have 2 manipulated variable (cooling water temperature and ratio of reactants) to control one variable (e.g. temperature) of the process?

Thanks,
Joseph

[Padmakar Katre]

a)I would like to know under what circumstances, the product concentration from the outlet of the reactor would be an important variabel to be controlled.

The general answer to a general question is the product outlet concentraion is controllerd inorder to maintain the reactor parameters like conversion and selectivity. The concentration of the outlet product control is important in order to optimize the reactor system. Some of the main issues which I feel which will always suggest you to control the outlet concentraions of the reactors are
1. In some reactions which are highly exothermic in nature and you need to maintain the conversion rates within limit.
2. The outlet concentarion of the reactor streams may have some limitations on the concentraion based on the flammability calculations which is function of the concentraion and which can lead to the post ignitions or the runaway reactions.
3. The outlet concentraions will decide the required production of the units.

b)Is it possible to have 2 manipulated variable (cooling water temperature and ratio of reactants) to control one variable (e.g. temperature) of the process?

Yes as per knowledge we can have. As we can not have much margins in the ratios of the reactants and that to the absolute feed flows of the individual reactants. In this event if your controller reaches the allowable limit and you don't get rid-off the problem i.e. still the temperature is not in control then we should think of the cooling water system. I think split-range control system will be helpful in this case where the TIC signal will be splitted and given to the Raio controller and the Flow controller of the cooling medium.
Hope you will get some idea. Waiting for your comments.

[Zauberberg]

I don't think it's a wise idea to manipulate two independent variables in order to control one single process parameter. As I have experienced in the past, this is poor control system design practice and it usually leads to unstable operation and - eventually - absence of proper control. The behavior of two independent control loops is such that it tries to "match" one with each other, causing fluctuations of controlled variable.

[Padmakar Katre]

I don't think it's a wise idea to manipulate two independent variables in order to control one single process parameter. As I have experienced in the past, this is poor control system design practice and it usually leads to unstable operation and - eventually - absence of proper control. The behavior of two independent control loops is such that it tries to "match" one with each other, causing fluctuations of controlled variable.

Dear Zauberberg,
I do agree with your opinion that having 2 independent control loops will lead to poor control but I think if we split the output signal of the reactor outlet temperature and give it to the ratio controller and the TIC or FIC of the cooling medium flow, this configuration should not be a problem. If we want monitor the outlet temperature say in case of an exothermic reaction we have to monitor the conversion rates and the selectivity inside the reactor which will result in proper temperature control now inthis case the conversion ans selectivity will also depend on the kinetics and the reactant feed concentraions so ultimately we arrive at the feed flow flow manipulations. But there would be constarints for the variation in either the absolute flows or the ratios of the reactants in feed once these values are at their minimum or maximum margins then you need another manipulated varaible to control the temperature in this case it may be the coolent flow or the coolent temperature.
If my post doesn't make any sense please ignore.

[Zauberberg]

Hello Padmakar,

On the contrary, I do not think your post is without engineering sense - I'm just trying to observe the system from operational point of view, as I have had opportunities in the past.

"Keep it simple" - that was (and it still is) the basic rule in design, optimization and debottlenecking of any process plant. You can feed your controller with as much data as you want (or as you consider to be required or optimal), but the output to the executive element in the control system - in my opinion - should be only one. That makes the system flexible, easy to understand and control. Otherwise, you might need a control system scientist in the place of panel operator.

Interaction between two independent control loops is rather complex and unpredictable, based on my experience. I can compare this with, for example, distillation tower having both reboiler temperature (hot utility flow) and tower top temperature (reflux flow) in automatic/cascade control. This will simply not work. Each of these controllers will react based on other control loop response, and they will start to catch with each other, leading to large fluctuations of BOTH controlled variables.

Thank you for clarifying your post and have a nice day. And happy birthday (it was yesterday?)


Best regards,

[Padmakar Katre]

Dear Zauberberg,
Thanks a lot for the birthday wishes and your reply.

[hoyoku]

Thanks for all your comments. I will think about it. Perhaps I would do a relative gain array analysis to study the interactions between the variables. I still need some time to digest what you all have spoken. Thanks.

Joseph

[hoyoku]

Hi there, does anybody have a diagram of how the split range control system is being drawn? I need a drawing of the tubular reactor control system. Thanks.

Joseph