Excluding components from the Fractionator

CADET Troubleshooting
1

Hi,

I’m currently checking out the capabilities of the Fractionator object within CADET-Process.

Is it possible to exclude components from the analysis? I’m working with a model where the first component is H+, but it does not really make much sense to include that component in the fraction analysis. My flow sheet includes an Outlet with the ComponentSystem that includes H+. Could a solution be to add another Outlet that excludes the H+ component, i.e. with a new ComponentSystem? Or is there a simpler solution to this problem?

Best regards
Kristian

2

Hey,
assuming you’re on dev branch, you can simply exclude components using the optimize_fractionation method.

from CADETProcess.fractionation import FractionationOptimizer
fractionation_optimizer = FractionationOptimizer()
fractionator = fractionation_optimizer.optimize_fractionation(
    simulation_results,
    components=['A', 'B', 'C'],
    purity_required=[0.95, 0.95, 0.95]
)
print(fractionator.performance)
_ = fractionator.plot_fraction_signal()

Here, you just specify the components you are interested in.

3

If you don’t directly use the optimizer, you can also use the slice_solution method directly to modify the final solution.

from CADETProcess.solution import slice_solution

solution_new = slice_solution(solution_original, components=['A', 'B'],...)
4

Hi Johannes,

Oh I see, this solution works for me. Thanks for a lot for the support.

I do however still get some kind of weird result in the fractionator.performance object. I’m getting a suspicious runtime warning:

“RuntimeWarning: divide by zero encountered in true_divide return self.mass / self.process.V_eluent”

It could indicate that self.process.V_eluent is in fact equal to 0. Do you have an idea why this is so? Following your documentation I added a “flow_sheet.add_eluent_inlet(‘elution’)” where ‘elution’ is an Inlet. I was guessing that the V_eluent should be available from this Inlet? The recovery also seem to be off at a very high value!

5

Would you mind sharing the scripts?

6

Hi Johannes,

I just tried to modify the solution as you describe here using the following code:

results = process_simulator.simulate(procss)
results_new = slice_solution(results.solution.column.outlet, components=['A', 'B', 'C'])
fractionator = Fractionator(results_new)
fractionator.add_fractionation_event('start_A', 0, opt_frac_times[0]*60, 'outlet')
fractionator.add_fractionation_event('end_A', -1, opt_frac_times[1]*60)
fractionator.performance.recovery

However, it seems that the Fractionator only accepts results of type “SimulationResults”. I’m guessing that slice solution will modify the type of the results object?

7

Hi Kristian,

the problem here is that Solution != SimulationResults.

  • The Solution is a wrapper around the np.array that represents the solution of a unit (e.g. at the unit inlet or outlet).
  • The SimulationResults contains all solution objects plus some meta data.

Good thing is that you can simply add the components flag to the __init__ of the Fractionator.

results = process_simulator.simulate(process)
fractionator = Fractionator(components=['A', 'B', 'C'])

Does this solve your problem?

By the way, for readability, please format code blocks with backticks (```)

e.g.

```
print(‘hello’)
```

will be rendered as:

print('hello')
8

Ok right, that works. Thanks for the excellent support as well as the tip with backticks :slight_smile:

1 Like