3 replies to this topic

[Olaniyi]

Hi fellow Che-resourcers

 

Has anyone on the forum work in a LNG plant where they do an online opimisation of the mixed refrigerant composition? I would be very interested in knowing how this is being done and how the interaction between the parameters (component compositions, compressor power and suction and discharge pressue, heavy and light MR flow rate etc) are being managed.

 

Thanks, happy to provide more details or make my questions more specific if need be.

 

Olaniyi

 

[Pilesar]

Mixed refrigerant composition changes because the refrigerant components are lost from the system in a manner that is not directly proportional to composition. Refrigerant loading is affected by weather and day/night temperature variations. The easiest implementation of online optimization consists of using plant data as inputs to a rigorous steady-state model in an open loop. This would give guidance to trained operators to implement the targeted values. Because the refrigerant load varies as well as the refrigerant composition, it is not practical to try to hold a precise optimum. Assuming the plant feed composition is steady, the operators can do a good job controlling the refrigerant composition by feel. Where an open loop online optimizer adds significant value is when it acts as a lighthouse for the operators to check that they are within reasonable range that is robust enough to handle process disturbances. You may find that much of this value can be captured by steady state simulaiton runs on a daily or even weekly basis. If you are convinced that higher frequency modelling adds financial value, consider using software tools specially designed for online optimization, such as SimSci's ROMeo product.

[Art Montemayor]

Olaniyi:

 

Most, or all, LNG processes are licensed.  If there is any question, problem, or complaint about the acceptability of the refrigerant(s) used in the process, it is the licensor's contractual responsibility to correct, change, modify, or "fix" the problem.

 

Why are you interested in getting involved in such a situation?  If you are the operator of a $multi-billion LNG facility, you would not be wise to tinker, experiment, or try to "fix" the refrigerant system unless you were ready to forfeit your contractual rights with the licensor or the process.  That means you take total and full responsibility for the results of your actions.

 

Is that what you are proposing or contemplating?

[Olaniyi]

Thanks Pilesar and Art

@Art

I dont actually plan to do any hardware tinkering - just contemplating how possible it would be to implement an automated mixed refrigerant composition optimiser, and what would be the best parameters to  look at. I work as a process engineer in an LNG facility and I'm trying to look at how the liquefaction loop can be optimised. Its a C3-MR process thus much of the available room for optimisation is in the MR loop. Obviously as a starting point I'm trying to look for solutions which do not require (much) capital - hence the consideration of the MR composition which can be changed fairly easily by venting/draining and make-up of refrigerant components. Like Pilesar mentioned I'm aware of many of the interactions in the MR loop, but was curious to see if anyone working on an LNG plant has actually done this - I dont see much evidence of this from my internet research (just one plant that i can see http://www.gastechno...ema_Che_Din.pdf and not much detail unfortunately).

 

Thanks

Olaniyi

Edited by Olaniyi, 27 March 2014 - 02:18 PM.