Problem with Merge and Splitter since switched to Tespy 9

[HaSchneider]

Since I switched to tespy 9 I have an issue I cant solve myself. I hope someone else could help me with that.

The here shown model raises the following error:

Simulation crashed due to an unexpected error:
unable to solve 1phase PY flash with Tmin=584.132, Tmax=3000 due to error: HSU_P_flash_singlephase_Brent could not find a solution because Smolar [193.579 J/mol/K] is above the maximum value of 184.8686532 J/mol/K

It works, if the "makeup" and "blowdown" components incl. their Merge and splitter are removed from the network.

from tespy.connections import Connection, Ref
from tespy.components import (
    Turbine, Source, Sink, Pump, 
    CycleCloser, SimpleHeatExchanger, Valve, Merge, 
    Splitter,    DropletSeparator,)
from tespy.networks import Network
needed_temperature=230
makeup_factor=0.05
Tamb=20
max_pressure=100
heat=1000E2
needed_pressure=28
h_superheating_max_pressure= 3617
main_pressure=40

#setup network
model= Network()
model.set_attr(iterinfo=False)
model.set_attr(T_unit='C', p_unit='bar', h_unit='kJ / kg')

# create components
boiler = SimpleHeatExchanger('steam boiler' , dissipative=False)
bpt = Turbine('back pressure turbine')
valve = Valve('controlvalve')
hex_heat_sink = SimpleHeatExchanger('hex heat sink', dissipative=False)
feed_pump= Pump('feedpump')
cycl=CycleCloser('CycleCloser')

makeup=Source("Make-up water")
blowdown= Sink("blowdown wastewater")
merge = Merge("Makeup water feed", num_in=2)
split= Splitter("remove wastewater")

#create connections:
c05 = Connection(cycl, 'out1', boiler, 'in1')
c04 = Connection(boiler, 'out1', bpt, 'in1', label='c04')
c03 = Connection(bpt, 'out1', valve, 'in1', label='c03')
c01= Connection(valve, 'out1', hex_heat_sink, 'in1', label='c01')
c1 = Connection(hex_heat_sink, 'out1', split, 'in1', label='c1')
c3 = Connection(split, 'out1', merge, 'in1', label='c3')
c4 = Connection(merge, 'out1', feed_pump, 'in1', label='c4')
c5 = Connection(feed_pump, 'out1', cycl, 'in1', label='c5')
model.add_conns(c05, c04, c03, 
                c01, c1,c5,)

muw = Connection(makeup, 'out1', merge, 'in2', label='muw')
wawa = Connection(split, 'out2', blowdown, 'in1')
model.add_conns( c3, c4,  muw, wawa)

#set attributes:
muw.set_attr(m=Ref(c04, makeup_factor, 0), 
                T=Tamb,
                fluid={"H2O": 1}, 
                p0=needed_pressure)

wawa.set_attr(m=Ref(c04, makeup_factor, 0))

boiler.set_attr(pr = 1, power_connector_location="inlet")
c04.set_attr(fluid={"H2O": 1},  h= h_superheating_max_pressure)
bpt.set_attr(eta_s = 0.85, )
c03.set_attr(p=main_pressure)
c01.set_attr(p = needed_pressure)
hex_heat_sink.set_attr(pr=1,Q=-heat, power_connector_location="outlet")
c1.set_attr(x=0)
feed_pump.set_attr(eta_s =0.95)
c5.set_attr(p=max_pressure)

model.solve('design')

[fwitte]

Hi @HaSchneider,

thank you for reporting this. This is a starting value issue, which may come from the fact, that the starting values are selected differently between the two versions. What you can do is to solve in two steps:

feed_pump.set_attr()  # do not set eta_s here!
c5.set_attr(p=max_pressure, h=Ref(c4, 1, 10))  # instead add a referenced h specification here!

model.solve('design')

# then switch specifications
feed_pump.set_attr(eta_s =0.95)
c5.set_attr(h=None)

model.solve('design')

I will adapt the convergence_check method for the Pump class, to force liquid only inlet and outlet state. That will resolve the issue.

Best

Francesco

[fwitte]

I fixed the issue on dev branch already (#708), I will make a release next week, since there is another bugfix coming from a different context.

[HaSchneider]

This actually solved the issue, but the next is coming right up:
Adding a Pipe in the network, leads to aborting of the calculation for low heat values (<1E5).

I know that the massflows are getting very low there, but without the pipe, I have no problem.

from tespy.connections import Connection, Ref
from tespy.components import (
    Turbine, Source, Sink, Pump, 
    CycleCloser, SimpleHeatExchanger, Valve, Merge, 
    Splitter, Pipe)
from tespy.networks import Network
needed_temperature=230
makeup_factor=0.05
Tamb=20
max_pressure=100
heat=1E4

needed_pressure=28
h_superheating_max_pressure= 3617
main_pressure=40

#setup network
model= Network()
model.set_attr(iterinfo=False)
model.set_attr(T_unit='C', p_unit='bar', h_unit='kJ / kg')

# create components
boiler = SimpleHeatExchanger('steam boiler' , dissipative=False)
bpt = Turbine('back pressure turbine')
valve = Valve('controlvalve')
hex_heat_sink = SimpleHeatExchanger('hex heat sink', dissipative=False)
feed_pump= Pump('feedpump')
cycl=CycleCloser('CycleCloser')

makeup=Source("Make-up water")
blowdown= Sink("blowdown wastewater")
merge = Merge("Makeup water feed", num_in=2)
split= Splitter("remove wastewater")

pipe_cold= Pipe('condensate pipe',dissipative=True)

#create connections:
c05 = Connection(cycl, 'out1', boiler, 'in1')
c04 = Connection(boiler, 'out1', bpt, 'in1', label='c04')
c03 = Connection(bpt, 'out1', valve, 'in1', label='c03')
c01= Connection(valve, 'out1', hex_heat_sink, 'in1', label='c01')
c1 = Connection(hex_heat_sink, 'out1', pipe_cold, 'in1', label='c1')
c2 = Connection(pipe_cold, 'out1', split, 'in1', label='c2')
c3 = Connection(split, 'out1', merge, 'in1', label='c3')
c4 = Connection(merge, 'out1', feed_pump, 'in1', label='c4')
c5 = Connection(feed_pump, 'out1', cycl, 'in1', label='c5')
model.add_conns(c05, c04, c03, 
                c01, c1,c5,c2,
                c3, c4, )

muw = Connection(makeup, 'out1', merge, 'in2', label='muw')
wawa = Connection(split, 'out2', blowdown, 'in1')
model.add_conns(  muw, wawa)

#set attributes:
muw.set_attr(m=Ref(c04, makeup_factor, 0), 
                T=Tamb,
                fluid={"H2O": 1}, 
                p0=needed_pressure)

wawa.set_attr(m=Ref(c04, makeup_factor, 0))

boiler.set_attr(pr = 1, power_connector_location="inlet")
c04.set_attr(fluid={"H2O": 1},  h= h_superheating_max_pressure)
bpt.set_attr(eta_s = 0.85, )
c03.set_attr(p=main_pressure)
c01.set_attr(p = needed_pressure)
hex_heat_sink.set_attr(pr=1,Q=-heat, power_connector_location="outlet")
c1.set_attr(x=0)
#feed_pump.set_attr(eta_s =0.85)
c5.set_attr(p=max_pressure,h=Ref(c4, 1, 10))
pipe_cold.set_attr(pr=0.95, 
        Tamb = Tamb,
        kA= 500,
        L=1000, D='var',  
        ks=4.57e-5)
model.solve('design')

The solver does not seem to make any progress, aborting calculation. Residual value is 1.47e+04. This frequently happens, if the solver pushes the fluid properties out of their feasible range.

[fwitte]

This time it is a numerical issue, which was present also in the previous versions of tespy (I tried with v0.8.3).

First issue is, that the target pipe diameter will be below the minimum possible (by default 0.01 m), you can change that. On top of that, I removed the kA specification and set the outflow temperature instead (essentially to be equal to the ambient temperature). You will see, that the resulting kA (with set Tamb the component will calculate the corresponding kA in postprocessing) is lower than your set value. Then the numerical "stabilization" (modifying the logarithmic temperature difference during iterations) will push the outflow temperature to somewhere impossible. Maybe there is a more intelligent way to handle the case of the fluid temperature change crossing the ambient temperature during iterations. I will have to check.

pipe_cold.D.set_attr(min_val=0.001)
pipe_cold.set_attr(pr=0.95,
        Tamb = Tamb,
        L=1000, D='var',
        ks=4.57e-5)
c2.set_attr(T=20.01)
model.solve('design')
pipe_cold.set_attr(kA=pipe_cold.kA.val)
c2.set_attr(T=None)
model.solve('design')

[HaSchneider]

Ah, ok, thank you. But as those massflow and diameters are not a realisitic scenario, in this case its more a theoretical issue.