Tutorial to CASE 3
1. Problem description
This case deals with a heat exchanger. A tracer experiment
was conducted in the “cold” circuit of that exchanger, normally fed
with water at ambient temperature. Nominal flow rate is 1.7 m3/h.
The tracer was 82Br as NH4Br, it was monitored at the outlet of the
exchanger only. Injection was assumed to be a very short pulse.
The tracer restitution curve has two peaks, which we interpret
as the result of a short-circuit somewhere in the exchanger. Our
objective is now to determine the corresponding flow rate.
2. Running the software
The outlet signal is in the case3.txt file in the code2 folder.
Select a Dirac delta as the inlet function and this file as the outlet
signal in the Setup menu. Visualise the data. Time is in seconds.
Data is not area-normalised this time, the RTD routine will do the
normalisation when the Solve/Run menu is invoked.
The obvious choice for the model is perfect mixers in parallel.
Examination of the curve suggests values for the time constants of
each branch (a few seconds for the short-circuit, about 200
seconds for the main flow). It also indicates the short-circuit should
be a small fraction of the total flow, maybe a few percents. The
shape of the curve also indicates that the J’s should not be large,
say about 5. This information allows to select reasonable initial
values for the parameters and to run the optimisation successfully.
The fit is however not excellent, perhaps because the inlet signal is
not quite a Dirac function.
In the general case, this model may prove a bit tricky to
optimise. It may be wise to find parameters for the main peak first,
using the perfect mixers in series, and to add the second branch
only then.
3. Analysis of results
o Determine the flow rate in the short-circuit and in the main
flow,
o Is it possible to determine the volume of the “cold” side of
the exchanger?
docsity.com
