Simple Steam Cracker Yield help needed based on feed qualities 1

[che6mw]

Hi,

I am currently undertaking a project of understanding and modelling an Ethylene cracker (with either an LPG or naphtha feedstock).

I've found data that has given me simple mass balances based on generic yield data for both naphtha (typical Europe/US?)and ethane (typical Middle East?) feeds.

I've a few problems as I progress:

1) Whilst i've a mass balance across a steam cracker using a naphtha feed this does not take in to account the naphtha feed quality. I've heard of companies using online analysis and software such as SPYRO to optimise their cracking. This is way beyond the level my project needs - are there some simple equations I can use that take naphtha qualities (I am guessing PIONA, distillation curve and density) and skew the product yields accordingly?

2) Some steam crackers use LPG instead of naphtha. Is this pure ethane, propane or butane, or mixes of these? Do steam crackers running LPG hold individual stocks of these and blend as required for whatever yield they are aiming for? Assuming steam crackers take different qualities of LPG are there again simple equations that would help me skew the yield of the cracker based on certain feed qualities?


Sorry this is a big meaty question but i've seen there are a few steam cracker process engineers floating around on here and can't find these questions asked previously.

Many thanks in advance,

Matt

 

[AndreChE]

A Stem Cracker can operate many different ways.

For naphtha crackers you should consider the cracking severity. The cracking yields depend on the furnaces severity. A real operational mode to control it is by COT Coil Outlet Temperature) and if the plant have Spyro, it can predict, based on products/feedstocks prices, the best operating point for the plant (the most profitable severity). The highest COT gives you the highest severity and the highest ethylene yield until a certain limit, after it decreases (cracking kinetics).

There are some ways to define severity. Usually we use P/E or P/M ratios (P/E = Propylene/Ethylene; P/M = Propylene / Methane) based on weight yields.

In my opinnion you should use for your calculations an assumed P/E severity. It is usual to use in Europe P/E from 0.45 (ethylene oriented) to 0.55 (propylene oriented). It depends on the company (cracker) economics and consumers. The Cracker I work operates, normally, at P/E from 0.46 to 0.48.

For Middle East (mainly ethane crackers) it is used the ethane conversion as an index of severity.

Please don't forget that material balances and products distribution are completely different from ethane or LPG to naphtha crackers.

Regards,
AndreChE

 

[che6mw]

Thanks AndreChE for your help.

I've read a few papers and have a material balance across a steam cracker for the following feedstocks:

ethane
propane
butane
naphtha

I've some data to show how a change in severity affects the yield when using naphtha. But how would I adjust the yield based on naphtha quality? are there simple correlations that show as paraffinicity (for instance) of my naphtha increases my yields change by x,y,z,etc???

 

[AndreChE]

No there aren't.

Normally with a PIONA and a software like Spyro you correlate a feedstock quality in 128 components (many of them pseudocomponents).

Based on that and on cracking conditions, Spyro calculates the yields of different products (the same 128 components).

Please don't forget that yields are fully dependent on the coils geometry in the cracking furnaces. This geometry is an input for the Spyro predictor.

What you are asking...is not so simple!

AndreChE

 

[che6mw]

thanks again. Are you aware of anywhere I could get such correlation data so I can model a very generalised steam cracker that takes in to account feedstock data?

 

[AndreChE]

No I'm not.

If you want, you can send me your data and I will run it for you at a given P/E severity.

Just be aware that I will simulate with our cracking coils which are Linde PyroCrack 4-2 design.

Regards,
AV

 

[shouvik]

Hi Matt,

LPG used as a cracker feedstock is generally a mix of butane and propane (most are from refinery) Normally crackers processing LPG do not blend the feedstock to acheive desired product yields.

My guess would be to use a composition range for the individual components - and try to develop the model based on kinetics data (if available).

Thanks.
S