11 replies to this topic

[HailToPitt]

Group, just carousing this forum, it really seems like everyone prefers using pre-packaged software, ASPEN, Plus, HYSYS among others. I think back to my days in school not long ago, and ASPENOne really was the way to get a whole design done quick, do a full price analysis, all that fun stuff. the other side of that is that solely relying on the software for design can lead to forgetting first principles.

 

While I was in school, my research was in simulation, particularly in a human body environment.  It involved deriving all the transfer equations and then making a Simulink model to run with different process parameters. This was then pushed up the chain to the good people at the medical institute who designed a prototype. Point being, I never used any kind of ChE oriented software, and instead built the model from the ground.

 

Flashing forward, I'm now engineer for an old plant that no one has done much analytic work on since the 1960s. As a manager first, and an engineer second, I lost a lot in the couple years I've been out of school, and wound up having to refresh on topics as they became important (bad engineer, I know).  I'm again using MATLAB and Simulink to build the model from the ground up.

 

So the main takeaway is something my mentor told me: ASPEN is a great way to get an answer, but if you don't understand the background of what's happening, you'll end up putting garbage in and getting garbage out. As a result, I find it's been rewarding to derive the model and build it, with the idea that I will be able to fact-check later. Does anyone share this sentiment? Does anyone use a programming and first principles approach as opposed to using the ASPEN building blocks?

 

Am I a ChE masochist for trying?

[Bobby Strain]

I, for one, totally disagree with you. After you have been in the business a few more years, you will gain the wisdom to understand why.

 

Bobby

[gegio1960]

I share Bobby view. You need to remember "first principles" to take control of input to and output from simulators but you shouldn't waste time to "reinvent the wheel". Of course, there are exceptions to the above....

[PaoloPemi]

I utilize regularly Matlab (as well as Excel and other tools) in union with a process library (PRODE PROPERTIES) for all thermodynamic calc's including block operations as columns, reactors etc.,

in my case this seems a reasonable compromise,

you do not need to reinvent the wheel (i.e.thermodynamics & complex operations)

while you are free to add your own procedures for specific design / optimization

Also for many properties PRODE exports derivatives and Matlab has the capability to integrate over a determined interval of time.

By the way from Matlab you can possibly set a link also with other simulators, if you have access to one you can investigate this possibility

Edited by PaoloPemi, 15 August 2013 - 01:31 AM.

[curious_cat]

Often equipment designers or people working on unique reactors etc. will use MATLAB etc. to create their unit models. 

 

ASPEN has its role and other codes theirs. They compliment each other rather than replace. 

 

No point trying to write your own flow-sheeting algorithm these days, but there's plenty of scope to optimize kinetics etc. in setups outside ASPEN. In particular, a majority of people on this forum seem from refining / oil & gas where the canned software is really well developed. For those in chemicals, pharma etc. we often lack this luxury and must develop lot of plug in models that will then interface into a commercial flow-sheeting code. 

[RockDock]

It's good to know the principles behind the process. A process simulator can either help you understand those principles or become a crutch you lean on in lieu of principles.

 

Aspen is great for non-kinetic standard applications. Outside of that, you will need something else.

[curious_cat]

What I really hate is Simulator "monkeys". "Aspen said so, hence it must be true" 

 

That sort of attitude leads to ruin. 

[gegio1960]

curious,

you're right... but I've obtained more and more right answers from the simulators than from some "experts" around.

[curious_cat]

curious,

you're right... but I've obtained more and more right answers from the simulators than from some "experts" around.

 

 

Yes, that's the other side of the coin. I hate the other school of thought too where just because someone has worked on a process for 30 years he can never be wrong. Laws of Physics be damned. 

 

I think a healthy dose of skepticism is what's the best remedy. Skepticism towards software and towards so called experts. 

[Erwin APRIANDI]

Garbage in <---> Garbage out

 

I think this not only applicable to process simulation, but it is apply to any task

even by using Matlab, Excell spreadsheet, or even calculator if you do not know the basic and whether the input is correct or not you will ended up to get Garbage out.

 

If the usage of simulation tools such as, AspenOne, Unisim, ProMax, Olga, Pipesim, Pipephase is reducing the requirement of people to undestand less on basic, I believe the company which are applying this is on the edge of a cliff and waiting to fall down.

 

Any big engineering company or even operator in my industry (Oil and Gas) have a strict rules in manually validating (random) the caculation results of the simulation tools, and no new version of simulation tools will be used unless it has been validated that the results of the simulation tools are correct.

 

And after that, an engineer who do the simulation is not only simulation operator but he/she is mostly already know what is happening inside the simulation tools (good both on the basic of chemical engineering and also well trained in simulation tools basic and operation) so in case of any abnormality on the simulation results he/she can justified it. before the results is actualy use for design. The role of simulation output nowadays is very crutial, since it has to mimic the reality in order to make a safe design.

 

Nevertheless, if the use of simulation tools can reduce the requirement of engineer to know the basic. You can start asking your self, why an engineer is required to do the simulations, why don't you hire an IT guys to do the simulation?

[marchem]

"While I was in school, my research was in simulation,..."

I did the same,

after, I continue to utilize Matlab, Mathcad and others,
these tools are indeed very good for analysis, modeling etc..

For process calc's I have PRODE PROPERTIES which
links directly with Matlab, Mathcad, Excel etc.

I have not to create thermo models,
databases for fluid properties etc.
and final results (for comparable applications)
are equivalent to those from process simulators.

According my experience these
numerical tools require
a somewhat different attitude
(compared with standard process simulators)
since you may need to define your own
models and validate procedures and results.

Many engineers today have little time
to complete projects and they prefer
to trust simulators results
with little data validation,
that, of course, may save a lot of time.

Of course I understand that point of view.

Edited by marchem, 30 August 2013 - 01:50 AM.

[HailToPitt]

What you say about old timers rings true everywhere I think. There are only a few engineers here and the fear I have is no one is ever open to a new way of thought.

The general consensus is that simulation and controls was only about laplace transforms (and who wants to sit around doing those all day) so the majority of that knowledge is shot. Then we wonder why no one knows how to tune a control valve.